Skip to main content
Internet Archive's 25th Anniversary Logo

Full text of "Amphioxus and the ancestry of the vertebrates"

See other formats


' ' ' 






: LD 


u - ***- - 


(Columbia hitbrrsitg Utolotjtcal Scries. 




By Henry Fairfield Osborn, Sc.D. Princeton. 


By Arthur Willey, B.Sc. Lond. Univ. 

3. FISHES, LIVING AND FOSSIL. An Introductory Study. 

By Bashford Dean, Ph.D. Columbia. 


By Edmund B. Wilson, Ph.D. J. H. U. 






'' - '"'To 











All rights reserved 


Nottoooti . 

J. S. Gushing & Co. Berwick & Smith. 

Boston, Mass., U.S.A. 







THIS volume originated in a course of University lec- 
tures prepared at my suggestion by the author. It seemed 
important that he should bring within the reach of 
students and of specialists among other groups, his own 
extensive observations upon Amphioxus and other remote 
ancestors of the Vertebrates, as well as the general litera- 
ture upon this group. While our detailed knowledge of 
the structure and habits of these animals has been rapidly 
increasing in recent years, it is still in the main very 
widely scattered in monographs and special papers. 

Probably no single group illustrates more beautifully 
the principles of transformism ; for the Protochordates in 
their embryonic development exhibit remarkable reminis- 
cences of past adaptations, and, in their adult develop- 
ment, the most varied present adaptations to pelagic, 
deep-sea, littoral, free-swimming, and sessile life. As 
Lankester has shown, the Ascidians alone give us a whole 
chapter in Darwinism. But degeneration and change of 
function constitute only one side of their history. In 


viii PREFACE. 

progressive development some of these types have come 
to so closely resemble, superficially, certain of the larger 
groups of Invertebrates, such as the Molluscs and Worms, 
that it is only at a comparatively recent date they have 
found their way out of these groups into the Protochor- 
data. Many of these misleading resemblances are now 
interpreted as parallels of structure springing from parallels 
in life habit, seen not only in the general body form, but 
in special organs, such as the breathing apparatus of the 
Ascidians and Molluscs. 

By the side of parallelisms are real invertebrate and 
vertebrate affinities ; so that the problem of resolving 
these various cases of original and acquired likeness in 
their bearing upon descent has become one of the most 
fascinating which modern Zoology affords. For example, 
among the real invertebrate ties of the Protochordates are 
the ciliated embryos of Balanoglossus and Amphioxus, 
the Tornaria larva and ciliated ectoderm of Balanoglossus. 
The nervous system of Balanoglossus presents both ver- 
tebrate and invertebrate characters ; the respiratory sys- 
tem is identical with that of Amphioxus, while in the 
embryonic development there are many resemblances inter 
se. In short, in Balanoglossus and the Ascidians the 
invertebrate type of structure, whether original or ac- 
quired, predominates. But in Amphioxus the balance is 
far on the other or vertebrate side of the scale, and this, 
with its resemblances to lower forms, gives us the con- 


necting link between Protochordate and Chordate organ- 
isation. Before entering into any of these discussions, 
the author has given a thorough systematic and structural 
treatment, especially of Amphioxus. 

This exquisite form, Amphioxus, is of almost world-wide 
distribution and has enjoyed the attention of every great 
zoologist for over half a century, yet the most recent 
studies upon it have been among the most productive of 
discovery. Its interest and value as an object of biologi- 
cal education has steadily increased with the knowledge 
that in contrast with all the related forms, it stands as 
a persistent specialised but not degenerate type, perhaps 
not far from the true ancestral line of the Vertebrates. 

H. F. O. 








Cranium and Sense-organs 17 


Atrial Cavity 22 

Viscera 24 

Ccelom 26 

Structure of Pharynx 27 

Evolution of the Thymus Gland 29 

Endostyle 31 

Branchial Bars 32 

Musculature 34 



INTERNAL ANATOMY (continued') 46 

Vascular System 46 

The Excretory System 55 

Development of the Atrial Cavity 75 

Comparison between the Excretory System of Amphioxus and 

that of the Annelids 78 

Nervous System 82 

NOTES ..... . 98 





EMBRYONIC DEVELOPMENT .............. I0 5 

Fertilisation and Segmentation of the Ovum ....... 105 

Gastrulation .................. IO 9 

Growth of Free-swimming Embryo .......... "3 

Development of Central Nervous System . . , ..... 118 

Origin of Mesoderm and Ccelom ..... .... 120 

Origin of the Notochord ... .... 124 

The Pneoral " Head-cavities " of Amphioxus ...... 126 

Enclostyle and Pigment Granules .......... I2 9 

LARVAL DEVELOPMENT ............... J 3 

Formation of Primary Gill-slits, etc .......... J 3 

Formation of Secondary Gill-slits . . ....... J 35 

Club-shaped Gland and Endostyle .......... !3 8 

Continued Migration of Primary Gill-slits ........ 139 

Peripharyngeal Bands .............. I 4 

Atrophy of First Primary Gill-slit and Club-shaped Gland, etc. 140 

The Adjustment of the Mouth, etc ........... ! 43 

Equalisation of the Gill-slits ........... ! 4 8 

Further Growth of Endostyle, etc. *49 

Development of Reproductive Organs ......... IS 1 


Larval Asymmetry .............. J 55 

Explanation of Asymmetry of Mouth and Gill-slits . . . 157 

Larval Asymmetry not Adaptive and not Advantageous . . . 161 


Nervus Branchialis Vagi .... ........ J 3 

Stomodceum, Hypophysis, and Gill-slits ........ l6 S 

Endostyle or Hypobranchial Groove . ....... l &7 

Peripharyngeal Ciliated Bands of Ammoccetes ..... 168 

Thyroid Gland ..... .... 169 

Morphology of Club-shaped Gland of Amphioxus . . . . 1 70 

Przeoral " Nephridium " of Hatschek ......... 1 7 2 

Ancestral Number of Gill-slits 



IV. THE ASCIDIANS ............... 


Test, Mantle, Atrium, Branchial Sac .......... 181 

Dorsal Lamina, Endostyle, and Peripharyngeal Band .... 183 

Visceral Anatomy ................ 186 

Nervous System and Hypophysis .......... 188 

Circulatory System ............... I9 1 

Renal Organs ................. 194 

Comparison between an Ascidian and Amphioxus ..... 194 


Segmentation and Gastrulation ........... 197 

Formation of Medullary Tube and Notochord ...... 198 

Origin of Mesoderm ............... 199 

Outgrowth of Tail ............... 201 

Formation of the Adhesive Papilla .......... 204 

Cerebral Vesicle and its Sense-organs ......... 204 

Comparison of Tunicate Eye with the Pineal Eye ..... 207 

Stomodceal and Atrial Involutions .......... 209 

Formation of Alimentary Canal and Hatching of Larva. . . 214 

Clavelina and Ciona ............... 214 


Vacuolization of the Notochord ........... 216 

Mesenchyme and Body-cavity ........... 217 

Pneoral Body-cavity and Prteoral Lobe ........ 218 

Body-cavity of an Ascidian and Coalom of Amphioxus . . . 220 

Fixation of the Ascidian Larva ........... 222 

Reopening of Neuropore; Degeneration of Cerebral Vesicle; 

Formation of Definitive Ganglion ........ 223 

Primary Topographical Relations and Change of Axis . . . 226 

Formation of Additional Branchial Stigmata ...... 229 

First Appearance of Musculature .......... 235 

Alimentary Canal and Pyloric Gland ......... 235 

Appendicularia ................ 236 

Abbreviated Ontogeny of Clavelina .......... 239 

NOTES ..................... 240 






External Features 242 

Nervous System and Gonads 244 

Metamerism 246 

Body-cavities; Proboscis-pore; Collar-pores 247 

Alimentary Canal 249 

Development; the Tornaria Larva 250 

The Larva of Asterias Vulgaris; Water-pores and Pnxoral 

Lobe 253 

Apical Plate of Tornaria 255 

Metamorphosis of Tornaria 256 

The Nemertines 256 




Anterior and Posterior Neurenteric Canals, and the Position of 

the Mouth in the Protochordates 274 




The Ascidian Hypophysis ... 287 


NOTES 291 


INDEX . . 311 


THE first zoologist to put forward, in a definite manner, 
the view of the existence of a direct relationship between 
Vertebrates and Invertebrates was the celebrated ETIENNE 

It would appear that without any previous zoological 
training, having been brought up as a botanist and 
mineralogist, he was appointed Professor of Vertebrate 
Zoology at the Museum of the Jardin des Plantes in the 
year 1793, being then twenty-one years old. His col- 
league as Professor of Invertebrate Zoology was the no 
less distinguished Lamarck. 

Saint-Hilaire's study of the comparative anatomy and 
osteology of the different groups of Vertebrates - - Fishes, 
Amphibians, Reptiles, Birds, and Mammals -- impressed 
him strongly with the conviction that, in spite of the 
many obvious contrasts existing between these animals, 
they are nevertheless essentially constructed upon the 
same plan, the same parts recurring in all the groups 
under a more or less altered form. Moreover, such 
observations as, for example, that the bones of a fish's 
skull can be more readily compared with the bones of an 
embryonic mammalian skull than with those of the adult, 
and that the bones of a bird's skull are separated in the 
young by sutures just as they are in the skull of a 
mammal, led him to frame his three great principles in 


terms of which the phenomena of animal organisation 
were to be, to a certain extent, explained. 

The three principles of Saint-Hilaire, each of which 
contains a large element of truth, were the following : 

1. The Theory of Analogues, according to which the 
same parts occur, in various grades of form and develop- 
ment, in all animals. 

2. The Principle of Connexions (Le principe des con- 
nexions), according to which the same parts always tend 
to occur in similar topographical relations. 

3. The Principle of the Correlation of Organs (Le 
principe du balancement des organes), according to which, 
cceteris paribus, the bulk of the animal body remains in 
a measure the same, and any given organ can only become 
enlarged or reduced according as another organ becomes 
reduced or enlarged. 

Having established these principles in his own mind 
from the exclusive study of the Vertebrates, the thought 
next occurred to him that probably they were capable of 
equal application to the rest of the animal kingdom, and 
he therefore undertook the task of identifying in the 
Insects the typical structural peculiarities of the Verte- 

According to his theory he would expect to find in the 
Insects, in some form or other, the same organs that 
occur in the Vertebrates. At the outset he was, as his 
successors have since been, confronted by the palpable 
fact that, while the longitudinal nerve-cord of the Insects 
lies next to the ventral surface of the body, the spinal 
cord of the Vertebrates lies below the dorsal surface. 
Accordingly he came to the conclusion which has since 
been strongly advocated by the upholders of the so-called 
"Annelid-theory," that the "back" and "belly" of an 


animal were gross conceptions of the ignorant and had 
no morphological meaning. These expressions merely indi- 
cated the position which an animal assumed in locomotion 
relative to the earth, and were in this sense convertible 
terms, since many invertebrate animals prefer to swim on 
their "backs," while some fishes also do the same, others 
again (flat-fishes, Pleuronectidae) swimming on their sides. 

The surfaces of the body in the respective groups having 
been thus reconciled, Saint-Hilaire proceeded to a detailed 
comparison between an insect and a vertebrate. The chiti- 
nous rings of an insect represent the vertebras of the higher 
animals. The viscera of an insect are thus enclosed within 
its vertebral column, and this condition is compared with 
what is found in turtles and tortoises where the carapace is 
fused with the vertebral column. It was necessary to con- 
clude, and Saint-Hilaire did not hesitate to do so, that the 
legs of insects were equivalent to the ribs of Vertebrates. 

It was not the intention of Saint-Hilaire to speculate 
concerning the ancestry of the Vertebrates, for this would 
have been impossible at the period in which he did his 
work, but he merely wished to demonstrate the truth of 
his principle of the unity of the plan of composition of the 
animal body. He had therefore no reason to be satisfied 
with having shown, as he believed, how the Insects could 
be regarded as possessing a structure essentially similar to 
that of the Vertebrates, but he had next to show how his 
principle could be applied to other groups, above all to the 
group of the Cephalopod Molluscs (squids, cuttle-fish, etc.). 
This happened in the year 1830, and it precipitated the 
celebrated and somewhat bitter dispute between the great 
Cuvier and Saint-Hilaire with regard to the question of 
" types." While Saint-Hilaire only recognised one uni- 
versal type, Cuvier arranged the different groups of animals 


under four entirely distinct types ; namely, Vertebrata, 
Mollusca, Articulata, and Radiata. Cuvier's system of 
classification remained in use for many years; in fact, until 
the progress of knowledge necessitated the adoption of a 
better one. 

For the first time, in 1864, the attempt was made by 
LEYDIG to grapple with the problem of the origin of the 
Vertebrates in the light of Darwin's Theory of Evolution 
(1858). Singular to say, although Leydig approached the 
subject from an entirely different point of view from that 
of Saint-Hilaire, yet he also attempted to find points of 
affinity between the highest Insects and the Vertebrates, 
and to identify the various subdivisions of the Vertebrate 
brain in the brain of the bee. 

Leydig and all those later authors who would derive the 
Vertebrates from an articulate ancestor, have started out 
with the a priori conviction that the segmentation of the 
body (metamerism) which is such a prominent feature (at 
least with regard to the musculature and skeleton) in 
fishes, and can be traced throughout the vertebrate series, 
especially in the embryonic stages, is morphologically 
identical with the familiar annulation or segmentation of 
the Articulates (Annelids, Arthropods). 

This is obviously a very natural assumption to make, but 
there is a large mass of facts which run counter to it, some 
of which will be referred to in the following pages. 

An unexpected light was thrown upon the problem of 
Vertebrate descent in 1866, when the Russian naturalist 
KOWALEVSKY published an account of his researches on 
the embryology of Amphioxus and the Ascidians. 

The Ascidians or Tunicates form a curious and in some 
respects well-defined group of animals, which used to be 
generally regarded as a subdivision of the Mollusca and as 


being closely related to the section of the bivalves or 
Lamellibranchiata. Kowalevsky, however, discovered that 
their embryonic development takes place on a plan so 
similar to that of Amphioxus as almost to amount to an 
identity. The development of the nervous and respiratory 
systems, and of the axial skeleton or notochord in the 
Ascidian embryo, as determined by Kowalevsky, showed 
in the clearest manner that the relationship of the Ascidians 
to Amphioxus, and through the latter to the Vertebrates, 
was an extraordinarily close one. 

Kowalevsky 's discovery of the chordate or sub-vertebrate 
character of the Ascidian larva, was considered by HAECKEL 
as affording a direct solution of the problem of the con- 
necting link between Vertebrates and Invertebrates. This 
was a somewhat extreme view to take of the matter, since 
Kowalevsky showed that the Ascidians could no longer be 
regarded as true Invertebrates. 

In 1875 the foundation of the Annelid theory of 
Vertebrate descent was laid independently by SEMPER and 
DOHRN ; and Kowalevsky's observations were explained 
away in favour of the new line of speculation. It was the 
discovery of the segmental origin of the excretory tubules 
of the Selachian (shark) kidney, made independently and 
simultaneously by SEMPER and BALFOUR, which may be 
said to have led to the definite framing of the Annelid 

Dohrn approached the subject from a different point of 
view. According to him, not only were the Vertebrates 
not descended from forms allied to the Ascidians and 
Amphioxus, but the latter were, by a process of almost 
infinite degeneration, derived or degenerated from the 

That the Ascidians are degenerate animals, to the 


extent that they have become adapted to a fixed habit of 
life, is of course obvious ; but that they have phyloge- 
netically undergone the immeasurable degeneration which 
was postulated by Dohrn, is a view which is entirely 
unjustified by facts. We shall now proceed to a presen- 
tation of some of these facts, devoting the first two 
chapters to the anatomy of Amphioxus, the third to the 
development of Amphioxus, the fourth to a brief sketch of 
the structure and development of the typical Ascidians, and 
the fifth to a consideration of the more abstruse relation- 
ships of the lower Vertebrates or Protochordates. 

The following classification of the forms more particu- 
larly dealt with may be of service : - 


Division i. HEMICHORDA (Balanoglossus, Cephalodiscus, 

and Rhabdopleura. See Chap. V.). 
Division 2. UROCHORDA (Ascidians). 
Division 3. CEPHALOCHORDA (Amphioxus). 




THE historical progress of our knowledge of Amphioxus 
has often been told, but for the sake of completeness it 
may be well to sketch its main outlines once more. 

It is interesting as being one of the few animals that 
were not known to Aristotle, having been described and 
figured for the first time in 1778 by the German zoologist 
PETER SIMON PALLAS. Pallas based his description on 
a specimen preserved in spirit, which had been sent to 
him from the coast of Cornwall; and as he confined him- 
self to the examination of the external form, he made 
what may appear to us the somewhat gross error of re- 
garding it as a Mollusc, a species of slug, and he accord- 
ingly named it Limax lanccolatns. He gives a perfectly 
recognisable figure of it, but was led astray by its flattened 
and pleated ventral surface, which might be construed 
into bearing a faint resemblance to a Molluscan "foot." 

This not very extensive knowledge of Amphioxus served 
the zoological world for nearly sixty years, until, in 1834, 
it was discovered for the second time in the Mediterra- 
nean, by the Italian naturalist, GABRIEL COSTA. Costa 
found it on the shores of Posilippo, in the Gulf of Naples, 
and was the first to make observations on the living ani- 
mal and to recognise its true nature. He thought at first 


that he had absolutely discovered it, but subsequently came 
across Pallas's description. He showed that it was ajis/i 
allied to the Cyclostomata, a group which includes the 
lampreys and hag-fishes. 

In his account of its habits he pointed out how sensitive 
it was to light, and although without apparent eyes, yet the 
light stimulated it to such an extent that it could by no 
means tolerate it. Costa mistook the curious tentacle-like 
processes or cirri, which form a circlet round the mouth 
(see Fig. i, p. 12), for respiratory filaments or branchiae, 
which suggested to him the name of Branchiostoma for the 
genus, the specific name given by him being Inbricnm, 
referring to the way in which it slips through the fingers 
with the rapidity of an electric spark when touched. 

WILLIAM YARRELL, in his History of British Fishes 
(1836), was the next to describe the remarkable creature 
and to give it the name Amphioxus, by which it has become 
so well known and which refers to the fact that it is pointed 
at both ends. Yarrell was also the first to describe the 
notocJiord or chorda dorsalis of Amphioxus as a cartilagi- 
nous vertebral column. 

Subsequently other observers had taken specimens of 
Amphioxus from various points, notably from the coast of 
Sweden, so that the attention of morphologists was at 
last definitely directed to the interesting form, and in 
1841 there were produced three independent memoirs on 
the anatomy of Amphioxus, which laid the foundation 
of our present knowledge. The authors of these memoirs 
Konigsberg, and JOHANNES MULLER of Berlin. The work 
of the last-named author is a masterpiece. With regard 
to the systematic position of Amphioxus, the outcome of 
all these researches was, that it was allied to the Cyclo- 


stomata, but, as Johannes Muller put it, differed from them 
to a greater extent than a fish differs from an Amphibian. 


In consequence of the extension of the firm, and at the 
same time elastic, notochord to the tip of the snout, 
Amphioxus possesses an extraordinary capacity for bur- 
rowing in the sand of the sea-shore or sea-bottom. If an 
individual be dropped from the hand on to a mound of 
wet sand which has just been dredged out of the water, 
it will burrow its way to the lowest depths of the sand- 
hillock in the twinkling of an eye. 

The frontispiece is designed to illustrate the chief 
positions in which Amphioxus may be observed. It is 
represented swimming, lying on the sand, and buried in 
the sand. 

Its usual modus vivcndi is to bury the whole of its 
body in the sand, leaving only the mouth with the ex- 
panded buccal cirri protruding. When obtained in this 
position in a glass jar a constant inflowing current of 
water in which food-particles are involved can be ob- 
served in the neighbourhood of the upstanding mouths. 

The food consists almost entirely of microscopic plants 
(Diatoms, Desmids, etc.) and vegetable debris. 

While passing through the pharynx the food becomes 
involved in the slimy secretion of a gland at the base of 
the pharynx known, as the endostyle or hypobranchial 
groove (cf. Figs. 2 and 3), and is thus held in the pharynx 
while the water with which it entered flows out through 
the gill-slits into the atrial chamber. The food is then 
carried through the intestine enveloped in a continuous 
cord of slime or mucus, which is kept in perpetual motion 


and rotation by the action of the cilia with which the 
epithelium of the alimentary canal is richly provided. 
After the digestible elements in the food have been dis- 
solved in the secretions of the intestinal wall the cord of 
slime with the attached faeces is duly ejected. 2 * 

The extreme shyness to a bright or sudden light which, 
as Costa observed, is manifested by Amphioxus, is prob- 
ably correlated with the presence of black pigment spots 
in the nerve-cord. If a lighted candle is carried into a 
dark room in which Amphioxus are being kept in glass 
jars, the excitement produced among the small fish is 

Occasionally it emerges from its favourite position in 
the sand, and after swimming about for some time it will 
sink to the bottom, and there recline for a longer or 
shorter period upon its side on the surface of the sand. 
When resting on the sand, it is unable to maintain its 
equilibrium in the same position as an ordinary fish would 
do, but invariably topples over on its side, indifferently on 
the right or left side. 3 In the higher fishes, including the 
lampreys, there is a special apparatus for controlling the 
equilibrium ; namely, the semicircular canals of the ear. 
There is nothing of the kind in Amphioxus, but in the 
Ascidian larva and in the Appendicularias there is, as 
we shall see, a structure situated in the floor of the brain 
known as the otolitli, which possibly exercises an equilib- 
rating influence. 

From what has been said above it follows that Amphi- 
oxus is an entirely passive feeder ; it does nothing in the 
way of biting, or even sucking, and has not to search far 
for its food, but merely takes what is brought in with the 

* This number and others which are scattered through the text refer to the 
Notes at the ends of the chapters. 


water which is drawn into the mouth by the powerful 
ciliary action of the cells lining the roof of the mouth and 
the wall of the pharynx. 

Speaking generally, Amphioxus is an inhabitant of 
shallow water; it is essentially a littoral form, and is apt 
to occur in the neighbourhood of any sandy shore. Its 
occurrence, however, is often curiously local, as shown by 
its behaviour at Messina. In the vicinity of Messina 
there are a couple of rather extensive salt-water pools, at 
some points of considerable depth, which, in the course of 
ages, have apparently been shut off from the adjacent sea 
by the formation of sandbanks. In the more northerly of 
these small lakes, lying almost at the extreme north- 
eastern point of Sicily, Amphioxus occurs in astonishing 
abundance ; while in the more southerly lake, which is 
connected with the former by a narrow artificial canal, it 
is entirely absent. Both of these lakes communicate 
by narrow outlets with the Straits of Messina, where, 
however, Amphioxus is somewhat rarely met with. In 
the Gulf of Naples it is extremely abundant ; while in 
Plymouth Sound, in the English Channel, it is compara- 
tively rare. On the coast of France it is said to grow to 
an unusually large size. It has been taken in greater or 
less numbers from many other localities in Europe, on 
the Atlantic and Pacific shores of North and South 
America, and from the shores of Australia, Japan, and 
Ceylon. Its geographical distribution may therefore be 
said to be pre-eminently world-wide, and, in fact, it is 
liable to turn up on any shore in the temperate and 
tropical regions. And yet with all this world-wide distri- 
bution there is only a single genus, with some eight 
species, 4 the different species being remarkably alike, 
differing slightly in the height of the dorsal fin and in 


the number of muscle-segments, the latter forming one 
of the chief diagnostic characters for a given species. 

The extensive geographical distribution of Amphioxus, 
combined with the fact that it is a shore-dweller and not 
a roving pelagic animal, and also with its remarkably 
constant features and, as a rule, trifling specific differ- 
ences, shows that we have to do with an extremely 
archaic form. 


A good idea of the external appearance an propor- 
tions of AuipJiioxiis lanceolatits can be obtained from the 
accompanying figure (Fig. i). Its actual length varies 


Fig. I. Amphioxus Lanceolatus from the left side, about twice natural size. 
(After LANKESTER.) The gormdic pouches are seen by transparency through the 
body-wall ; the atrium is expanded so that its floor projects below the metapleural 
fold ; the fin-chambers of the ventral fin are indicated between atriopore and anus. 
The dark spot at the base of the fifty-second myotome represents the anus. 

from about four to as much as eight centimetres. In 
the fresh condition it is semi-transparent, so that some 
of the internal organs can be seen through the skin, which 
is often iridescent. 

The figure shows the pointed extremities of the body 
and the circlet of tentacles or buccal cirri round the mar- 
gin of the mouth, or more accurately, the oral hood, 
because the mouth proper is covered over by a hood-like 
fold of the integument, from the margin of which these 
processes grow out. Extending from near the anterior 
extremity of the body to the posterior end are seen some 
sixty-two oblique parallel lines, each bent upon itself in 


such a way as to form two sides of a triangle, the apex 
of which is directed forwards. These are the partitions 
or septa which divide the longitudinal muscles of the 
body into a series of separate muscle-chambers or vivu- 
touics. In virtue of the longitudinal muscles being broken 
up, so to speak, into a great number of segments, the 
animal is enabled to swim rapidly with a serpentine 
motion. In the remarkable pelagic animal, Sagitta, where 
the muscles are not segmented, this motion is impossible, 
and instead, it darts forward by sudden and spasmodic 
jerkings of its tail. 

In Amphioxus, the tail or post-anal region of the body 
is very much reduced, and the muscle-segments of the 
trunk therefore constitute its only means of locomotion, 
there being no muscular fins. Beyond the muscle-plates, 
both in front and behind, the notocJiord, which forms the 
axial skeleton of the body, is seen to extend to the anterior 
and posterior extremities. The extension of the notochord 
beyond the anterior limit of the dorsal nerve-tube is a very 
exceptional condition, and has led to the creation of a 
special order for the reception of Amphioxus ; namely, the 

The oval structures seen lying below the muscle-plates 
in Fig. i are the reproductive organs, male or female as 
the case may be. Instead of being represented by a single 
genital gland on each side of the body as they are in the 
higher fishes and Vertebrates generally, they consist here 
of some twenty-six pairs of perfectly distinct chambers, 
occurring in correspondence with the muscle-segments or 
myotomes of the region to which they belong, and extend- 
ing from the tenth to the thirty-fifth myotome inclusive. 
These chambers are known as the gonadic poucJics. (See 
Fig. 2.) 


About two-thirds of the way from the front end of the 
body there is a comparatively large aperture in the mid- 
ventral line. It is the excurrent orifice of a spacious 
cavity which surrounds to a large extent the internal 
organs, including above all the pharynx, and is known as 
the atrial chamber, or simply atrium, while its opening to 
the exterior is the atriopore. 

The anus or outlet of the digestive tract occurs near the 
posterior end of the body ; it does not lie in the mid- 
ventral line, but high up on the left side. At its first 
appearance in the young embryo, the anus does lie ap- 
proximately in the mid-ventral line (cf. Fig. 64, p. 117), 
but as soon as the caudal fin begins to develop, it is pushed 
on to one side, always the left, and so attains its final 
position. A similar displacement of the cloacal aperture 
occurs in the Dipnoan fish Protopterus, where, however, 
the direction of displacement is not constant, the aperture 
lying now to the right, now to the left, of the middle line. 
Again, in the tadpoles of certain Batrachians the cloacal 
aperture is displaced to the right of the middle line.* (Cf. 
Fig. 8.) The fact of the displacement of openings by the 

* The asymmetrical position of the cloacal aperture of certain Batrachian 
tadpoles has been systematically worked out by BOULENGER. In tadpoles 
of the genera Rana and Hyla, the cloacal aperture is dextral, while in the 
Toads and Pelobatoids it is median. (See G. A. BOULENGER, A Synopsis 
of the Tadpoles of the European Batrachians. Proc. Zool. Soc. London, 
1891. pp. 593-627. Plates 45-47.) 

In Rana the cloacal aperture may occasionally occur in a median position 
as a variation. (WiLLEY, Note on the position of the cloacal aperture in 
certain Batrachian tadpoles. Transactions New York Acad. of Sciences, Vol. 
XII. 1893. pp. 242-245.) My attention to the previous literature on this 
subject was kindly drawn by Mr. G. A. Boulenger. 

Since writing the above my attention has been called to the following 
paper by Professor BURT G. WILDER, Lateral Position of the Vent in Am- 
phioxns [Branchiostoma] and in the Larva: of Rana Pipiens [Catesbiana]. 
Proc. Amer. Assoc. Adv. Sc. XXII. 1873. pp. 275-300. 


differential growth of neighbouring structures is a very ^ 
oiis one, and should be borne in mind. It will have a specia 
significance when we come to consider the development. 

There are no paired muscular fins in Amphioxus, but 
running along the whole length of the back is a median 
ridge which is called the dorsal fin. It extends round the 
front end of the body, where it becomes continuous with 
the right half of the oral hood. (Cf. Fig. 9.) Posteriorly 
it becomes enlarged to form the tail expansion or caudal 
fin, and is continued round the hinder extremity of the 
body past the anus as far as the atriopore. Along the 
back, this continuous fin is supported by a series of gelat- 
inous fin-rays, each of which lies in a chamber of its own. 
The fin-rays, whose number may exceed 250, do not extend 
to the extreme anterior and posterior ends of the body. 
The ventral portion of the fin in the region between atrio- 
pore and anus is supported by a similar series of fin-rays, 
but there are two of them placed side by side in each com- 
partment. In other words, the fin-rays of the ventral fin 
are paired. 

Amphioxus, like most fishes, is laterally compressed so 
that a transverse section through the body in front of 
the atriopore is found to have the form of an equal-sided 
spherical triangle, the base of which consists of the floor 
of the atrial chamber. At each of the basal angles of 
the triangle there is a fold of the integument containing 
a cavity (Fig. 2). This is the metapleural fold 1 which 
stretches on each side of the body from the region of 
the mouth to slightly beyond the atriopore. (Cf. Fig. i.) 
The cavity in the folds is the metapleural lymph-space. 
The apex of the triangular cross-section is formed by 
one of the dorsal fin-chambers enclosing a lymph-space 
into which a fin-ray is projecting. 





Fig. 2. Diagrammatic transverse section through pharyngeal region of female 
Amphioxus. (After LANKESTER and BOVERI, from R. Hertwig's Lehrbuch d. 

at. Atrial cavity, c. Dorsal coelom, separated from atrial cavity by the double- 
layered membrane known as the ligamentum denticulatum. ch. Notochord. 
d.n. Dorsal spinal nerve, e. Endostyle, below which is the endostylar ccelom con- 
taining the branchial artery, f. Fin-ray of dorsal fin. g. Gonadic pouch contain- 
ing ova. h.v. Hepatic vein lying in the narrow coelomic space which surrounds 
/, the liver or hepatic ccecum. La. Left aorta separated from the right aorta by the 
hyperpharyngeal (epibranchial) groove. ly. Lymph-space. mp. Metapleur. 
my. Longitudinal muscles of myotomes ; over against the dorsal coelom these 
muscles are arranged vertically, and form the rectus abdominis of Schneider. 
n.t. Spinal cord. p. Pharynx, r. Excretory tubule, t.m. Transverse or subatrial 
muscles, v.n. Ventral (motor) spinal nerve, the fibres of which have the appear- 
ance of passing directly into the muscle-fibres. 

N.B. The connective tissue (cutis, notochordal sheath, ccelomic epithelium, 
etc.) is indicated by the black lines 


In young transparent individuals, such as that of which 
the anterior portion is represented in Fig. 3, the pharynx, 
or first division of the digestive tract, into which the 
mouth leads directly, can be seen through the body-wall, 
and it is found to be perforated on each side by a great 
number of elongated vertical slits, whose number varies 
with the age of the individual, but may eventually attain 
the astonishing figure of 180 pairs. They are the gill- 
clefts opening from the pharynx into the atrial chamber. 
In the living Amphioxus an almost continuous stream of 
water is being drawn through the mouth into the pharynx 
for purposes of respiration and nourishment, then pass- 
ing out of the pharynx, by way of the gill-clefts, into 
the atrial chamber and thence to the exterior through the 

Cranium and Sense-organs. 

Besides lacking differentiated lateral fins, Amphioxus 
differs fundamentally from the higher Vertebrates in the 
absence of a cranium, of paired eyes, and paired or un- 
paired auditory organs. 

On account of the absence of a cartilaginous cranium 
it has been placed by itself in a separate division, the 
Acrania, in contrast to all the other Vertebrates proper, 
from the Cyclostomata upwards, which all possess a 
cranium of one sort or another and are hence known as 
the craniate Vertebrates or Craniota. In Amphioxus the 
only cartilage in the head-region consists of a ring lying 
round the margin of the oral hood at the base of ^ie 
buccal cirri. It is formed of separate pieces correspond^ 
ing to the number of the cirri, and each piece sends up a 
process into its adjacent cirrus, so that the latter is pro- 
vided with a stiff skeletal axis (Figs. 3 and 4). These are 



the buccal cartilages. As pointed out by Johannes Miiller, 
they are not to be compared with the jaw-apparatus, nor 

to the hyoid or tongue- 
bone of the jaw-bearing 
Vertebrates, but they 
belong to the same cate- 
gory as the mouth-carti- 
lages of the Cyclostome 
fishes (which possess a 
hyoid cartilage in addi- 
d.s end tion) and the labial car- 

Fig. 3. Anterior portion of body of young tllagCS of Selachians 
transparent individual. (After J. MtJLLER, / i i \ 
slightly altered.) 1Kb > 

ch. Notochord. ci. Buccal cirri, e. Eye- The absence of paired 
spot. end. Endostyle. f.r. Fin-rays lying in 

the fin-chambers. g.s. Gill-slits ; the skeletal CyCS and of any kind of 
rods of the gill-bars are indicated by black lines. audit hag b 

nt. Spinal cord, with pigment granules near its J 

base. r.a. Downgrowth from right aorta lying mentioned above. There 
to the right of vel. the velum ; with velar ten- . 

tacles projecting back into pharynx, w.o. Rad- 1S > however, a median 
erorgan; ciliated epithelial tracts on inner wh j ch consists of a 

surface of oral hood. J 

comparatively large un- 
paired pigment spot lying at the anterior extremity of the 
dorsal nerve-tube.* A row of 
similar, but much smaller, 
masses of pigment lie along 
the floor of the spinal canal, 
commencing some distance 
behind the eye (Fig. 3). 

Immediately above and be- Fig. 4. Buccal cartilages of Am- 
hind the eye-spot is a small P hio * US ' ( ^ fter i MUL , LER ') The 

basal pieces he end to end in the mar- 
pit in the body-wall reaching gin of oral hood, and each basal piece 

sends up an axial process into the 
from the OUter Surface of the corresponding buccal cirrus. 

* The eye-spot has been observed to be sometimes broken up into two 
pigment masses. (See AYERS, No. 105 bibliog.) 


body to the anterior wall of 
the brain. This is known 
as Kollikcr's olfactory fit, 
after its discoverer. The 
cells which line its walls 
carry long vibratile cilia, and 
it possibly subserves in some 
degree an olfactory func- 
tion. In the larva the cavity 
of the brain opens into the 
base of the olfactory pit by 
a pore known as the nenro- 
pore, which we shall consider 
later. In the adult this 
pore becomes closed, but 
the base of the olfactory pit 
appears to remain connected 
with the roof of the brain 
by a solid stalk. The olfac- 
tory pit, like the anal open- 
ing, lies asymmetrically on 
the left side of the body 
(Fig. 5). It is forced to one 
side in the course of the 
development consequent on 
the formation of the fin-like 
expansion of the integument 
in this region, which, as we 
have seen, is nothing more 
than the cephalic continua- 
tion of the dorsal fin. 

The mouth of Amphioxus 
would seem to be well 

Fig. 5. Transverse section through 
region of olfactory pit. (After LAN- 


The olfactory pit is seen as an ecto- 
dermic involution on the left side in con- 
tact with the wall of b, the cerebral vesicle. 
c/i. Notochord. / Lymph-space of ce- 
phalic portion of dorsal fin. r.h. and l./i 
Right and left portions of oral hood. 
my. Muscles of first myotome ; outside of 
the muscles is the myocoslic lymph-space 
of first myotome ; inside of the muscles 
is the apex of the myocrelic lymph-space 
of the second myotome. . Cranial 
nerve (second pair). 

N.B. The dotted shading represents 
the thickened gelatinous connective tissue 
of the head-region in which irregular 
lymph-spaces occur. 



guarded against the intrusion of noxious substances. 
Everything entering the mouth has to pass through a 
vestibule richly provided with sensitive epithelial cells. 
This vestibule consists of the oral hood with its marginal 
cirri, at the back of which lies the definite oral opening or 
velum, as it was called by HUXLEY on account of its 
resemblance to a similar structure in the young lamprey 
(Ammoccetes). (Cf. Fig. 3.) In the adult the velum 
carries twelve tentacles of its own, the velar tentacles, 
which are not to be confused with the buccal cirri of the 
oral hood. The velar tentacles project in a backward 
direction freely into the pharynx. 

A B 

Fig. 6. A. Portion of a buccal cirrus to show groups of sense-cells. 
B. Isolated ceLls of the skin ; two columnar sense-cells carrying a sensory hair, 
and one cylindrical epidermic cell with striated cuticular border. (After LAN- 


Groups of sense-cells occur on the side of the buccal 
cirri at intervals (Fig. 6). Some of these cells bear a 
vibratile cilium at their free ends, and others bear stiff 
hairs. Both kinds of cells are mingled in the same group. 



Similar groups of sensory cells occur on the margin of the 
velum and its tentacles (Fig. 7). It may be noted, in 
anticipation, that the velum is derived directly from the 
mouth of the larva, which 
becomes secondarily hid- 
den from superficial view 
by the overgrowth of the 
oral hood. 

According to LANGER- 
HANS, similar cells to 
those mentioned above, 
carrying stiff sensory 
hairs, are scattered dif- 
fusely all OVer the exter- Fig. 7. Velumof Amphioxus seen from 

nal surface of the body. the ide of the pharynx " (After LANKES - 


v.sp. Sphincter muscle of velum, v.t. Velar 
tentacles lying across the oral opening. 

(Cf. Fig. 6 B.) But a 
concentration of sense- 
organs comparable to the lateral line of the higher fishes 
is apparently absent. 7 

A remarkable structure which seems to combine the 
properties of gland and sense-organ occurs on the under 
surface of the oral hood. It consists of a patch of 
modified epithelium drawn out into finger-shaped epi- 
thelial tracts, the cells of which carry long cilia. (See 
Fig- 3-) It was discovered and accurately described by 
Johannes Muller, who called it the "Raderorgan" on ac- 
count of the resemblance of its ciliary movements to those 
of the wheel-apparatus of a Rotifer. The result of the 
combined action of the cilia is to cause a flow of water 
into the pharynx. In connection with the Raderorgan 
must be mentioned a special depression forming a peculiar 
sense-organ (Geschmacksorgan) lying against the right 
side of the notochord, known as the groove of HatscJick. 


At rial Cavity. 

In making a dissection of a frog or a fish, as soon as the 
body-wall is cut through, we find ourselves groping about 
in a large cavity in which the viscera lie. This is the 
body-cavity or peritoneal cavity, or, again, the ccelom. 

If we slit open the ventral body-wall of Amphioxus, we 
discover what appears to be an exactly similar cavity. It 
is, however, not the ccelomic cavity, but the peribranchial 
or atrial cavity, into which the pharyngeal gill-slits open. 
The older anatomists, including Johannes Miiller, regarded 
it as the true body-cavity, and the latter author was forced 
to the conclusion that Amphioxus differed fundamentally 
from all the other Vertebrates in that the gill-slits opened 
into the peritoneal cavity. Although that condition of 
things was hard to imagine, yet it seemed to be obviously 
the case, since the reproductive organs appeared to lie in 
the same cavity, and it went without saying that a cavity 
containing the gonads could only be the peritoneal cavity. 
In reality, the gonads do not lie in this cavity ; they only 
project into it and lie in a space of their own which is 
separated from the atrial cavity by a double-layered mem- 
brane. (Cf. Fig. 2.) 

HUXLEY threw some light on the matter in 1874, when 
he compared the atrial or peripharyngeal cavity of Amphi- 
oxus to the opercular cavity which surrounds the gills of 
the tadpoles of the frog and tailless Amphibia generally. 
In the case of the tadpole, as is well known, there are some 
four pairs of gill-slits which open at first directly to the 
exterior. Subsequently an opercular fold grows backwards 
over them as in fishes, but with this difference, that in the 


frog-tadpole the fold of one side becomes continuous ven- 
trally with that of the other, so that in effect we have one 
large semicircular fold covering over the gill-slits. Event- 
ually the hinder free margin of the fold undergoes con- 
crescence with the body-wall, so that a single peribranchial 
cavity is formed about the gills. This cavity is closed all 
round except at one point, usually on the left side, but 
sometimes in the mid-ventral line, where it remains open 
as i\\Q porus brancJiialis, or so-called spiraculum. 

This comparison of Hux- 
ley's was extremely well 
taken, and although the two 
cavities, namely, the peri- 
branchial cavity of the frog- 
larva and the atrial chamber 
of Amphioxus, are probably 
by no means homologous, or 
genetically related to each 
other, still the close analogy 
that exists between them is 
most instructive, and yet, 
singular to say, it did not 
lead Huxley to a correct Fig _ 8 _ Tadpole of Frog (Ranacla . 
interpretation of the atrial mata ^ from ventral side - (Original.) 

cl. Dextrally placed cloacal aperture. 

chamber. 5 m. Mouth, sp. spiraculum; the dotted 

T , line indicates the extent of the opercular 

Its true nature was at chamber , Tail> 
length established by ROLPH 

in 1876. By comparing his own observations on the adult 
with those of Kowalevsky on the larva, Rolph came to the 
conclusion that the atrial cavity of Amphioxus originated 
by the growth of two folds of the body-wall over the gill- 
slits on each side, and by their subsequent fusion in the 
mid-ventral line except at one point, which remained open 


as the atriopore. Although the details in the formation of 
the atrium are not exactly such as they were supposed to 
be by Rolph (see below), yet the end-result is virtually the 
same, and his work marks a distinct advance in our knowl- 
edge of the structure of Amphioxus, by showing that the 
epithelium lining the walls of the atrial chamber is not 
peritoneal, but is derived by a process of in-folding, from 
the ectodermic covering of the surface of the body. In 
other words, the atrial cavity, like the opercular cavity of 
the Amphibian tadpole, is lined by ectoderm. 


A bird's-eye view of the internal organs, as exposed by 
cutting the animal open ventrally by incisions extending 
forwards and backwards from the atriopore, is shown in 
Fig. 9. First and foremost, our attention is arrested by 
the relatively enormous pharynx occupying more than half 
the length of the body, with its right and left perforated 
walls and parallel gill-bars abutting at the mid-ventral line 
on the endostyle. 

The alimentary canal is seen in the dissection to have a 
perfectly straight course between mouth and anus, with 
no windings whatever. Growing out ventrally from what 
may be termed the pyloric region of the intestine, a short 
distance behind the pharynx and in front of the atriopore, 
there is a large diverticulum ending blindly in front, which 
in the adult lies for the greater part of its extent applied 
against the right wall of the pharynx (Fig. 9). This is 
the so-called Jiepatic ccecuni, corresponding to the liver of 
higher forms. The permanent condition of the liver in 
Amphioxus is comparable to its embryonic condition in the 
Vertebrates, where it attains a much more complicated 
structure in the older stages by subsequent branching and 


anastomosing of the branch- 
es, etc. It is essentially a 
median ventral outgrowth 
from the intestine, and its 
lying on one side of the 
pharynx in Amphioxus is 
only a secondary topographi- 
cal necessity.* 

Attached to the lateral 
muscular body-wall on each 
side are the gonadic pouches, 
which project into the cavity 
of the atrium. (Cf. Fig. 2.) 
Their number, which is usu- 
ally twenty-six pairs, varies 
slightly, and sometimes there 
are more on one side than 
on the other, as in Fig. 9. 

The atrial cavity does not 
end at the atriopore, but is 
continued beyond it as a 
blind sac lying to the right 
of the intestine, and reach- 
ing back nearly as far as the 
anus. In Fig. 9 the position 
of this post-atrioporal exten- 
sion of the atrium is indi- 
cated by means of a dotted 

Finally, in Fig. 9, the anus 
is seen lying to the left of 

* The coecum is held in position 
turn denticulatum. 




Fig. 9. Amphioxus dissected from 
the ventral side. (After RATHKE, slightly 

m. Entrance to mouth with the buccal 
cirri lying over it. /. Pharynx, e. Endo- 
style. /. Hepatic caecum, g. Gonadic 
pouches, at. Position of atriopore ; the 
post-atrioporal extension of the atrium is 
indicated by the dotted line passing over 
to the right side of i, the intestine, an. 

N.B. Note absence of differentiated 

by cord-like attachments to the ligamen- 



the caudal fin, and the right margin of the oral hood is 
shown to be continued round the front end of the body 
into the cephalic expansion of the dorsal fin. 


The question now arises : if the atrial cavity is no: 
the true body-cavity, what has become of the latter ? In 
order to determine this point, it is necessary to have 
recourse to transverse sections through the body, such as 
the one represented in Fig. 2, which is taken through the 
middle of the pharyngeal region. In a section like this, 
the work of tracing the limits of the atrial cavity is often 
greatly facilitated by the presence of a rich brown pigment 
in the epithelium lining its walls. We find, accordingly, 
that the atrial cavity has extended itself at the expense of 
the coelom, and has reduced the latter, in the main, to a 
small space on either side of the dorsal aorta, the aorta 
being double in this region (Fig. 2). This portion of 
the coelom is sometimes spoken of as the snpra-pharyngeal 
ccelom, and sometimes as the snbcJwrdal coelom, since it lies 
dorsal to the pharynx on the one hand, and below the noto- 
chord on the other. Other fragments, so to speak, of the 
coelom are found accompanying some of the branchial bars, 
namely, every alternate one ; and another portion occurs 
below the endostyle. (See Fig. 13.) The hepatic coecum 
is also surrounded by a division of the coelom, but its 
cavity is reduced to a minimum, and the same applies to 
the ccelom surrounding the intestine immediately behind 
the pharynx. Behind the atriopore, as we have seen, the 
atrial cavity is confined to the right side, so that on the 
left side of the intestine in this region the coelom presents 
its original proportions. 


Structure of Pharynx. 

We have already had occasion to mention the fact that 
the wall of the pharynx on each side is perforated by a 
great number of vertically elongated slit-like apertures 
the gill-clefts. In the middle region of the pharynx the 
gill-slits stretch almost from the roof to the base of the 
pharynx, but in front and behind they gradually become 
much lower in vertical height (Fig. 10). In the fully 


Fig. 10. Anterior portion of right wall of pharynx, to show arrangement of 
skeletal rods. (After J. MULLER.) 

f. Endostyle. e.c. Endostylar coelom. p.b. Skeletal rod of primary gill-bar. 
t.b. Skeletal rod of tongue-bar, sy. Cross-bars or synapticula. 

N.B. A simple gill-slit undivided by a tongue-bar should have been inserted 
in the figure in front of the first double slit. J. Miiller failed to observe this. 

expanded condition the gill-slits are nearly vertical, as in 
Fig. 10, but by the contraction of the transverse muscles, 
which lie in the floor of the atrium, they are often found 
to be directed very obliquely backwards, and this is the 
condition in which they almost invariably occur in pre- 
served specimens. That is the reason why so many of 
the bars are involved in a single transverse section. (Cf. 
Fig. 2.) On account of the prodigious extent to which 


the pharynx is perforated by the gill-clefts, it is necessary 
for it to have some sort of skeletal support to prevent it 
from collapsing. This is effected by a series of stiff gelat- 
inous rods which lie in the walls bounding the gill-clefts. 
These rods have the consistency of chitin, the material 
that forms the exoskeleton of insects, and are insoluble 
in caustic potash. The portion of the pharyngeal wall 
which lies between any two gill-slits is called a gill-bar. 

It will be seen at once in Fig. 10 that there are two 
kinds of skeletal rods differing in the behaviour of their 
lower extremities. Dorsally the rods arch over into one 
another, but ventrally they are independent, and every 
alternate rod is bifurcated, while the somewhat shorter 
intermediate rods end plainly. The forked rods form the 
skeletal support of the primary gill-bars, while the inter- 
mediate simple rods support the secondary gill-bars, or 
tongue-bars, as they are usually called. The primary bars 
constitute the walls of the primary gill-clefts. The latter, 
at their first origin, appear as simple oval openings in 
the wall of the pharynx. Later on the simple opening 
becomes divided into two by the gradual dipping down- 
wards of its dorsal margin until it meets and fuses with 
the ventral margin. In this way is the tongue-bar formed 
and the gill-slit doubled. (Cf. Fig. 11.) The statement 
which was made above, therefore, that there could be as 
many as 180 openings on each side of the pharynx, signified 
that there might be some ninety pairs of primary gill-clefts. 

Eventually the gill-slits become still further subdivided, 
though not so obviously, by the formation of small cross- 
bars which pass over from one primary bar to another, 
skipping over the tongue-bar, although eventually fusing 
with the skeletal axis of the latter on their inner faces 
(Fig. 10). 


2 9 

Evolution of tJic Tliymns Gland. 

Tongue-bars, like those occurring in the gill-slits of 
Amphioxus, are only known otherwise to occur in the 
remarkable worm-like creature, Balanoglossus. In the 
higher Vertebrates they appear to be entirely absent, but 
in the course of the development of the higher forms 
there is a structure which arises from the dorsal wall 
of the gill-slits which may very well be the homologue 
of the tongue-bars of Amphioxus. This structure is the 

olf vel ph.b nph 





Fig. it. Anterior region of young Amphioxus from left side. (After WlLLEY ; 
the renal tubules inserted after BOVERI.) 

at. Atrium, ci. Buccal cirri, ch. Notochord. d.f. Dorsal fin-chambers, e. Eye- 
spot. end. Endostyle. hep. Outgrowing coecum ; the index line passes through 
one of J. Miiller's renal papillae. met. Metapleural fold. nph. Nephridia or renal 
tubules, nt. Spinal cord. olf. Olfactory pit. ph.b. Peripharyngeal ciliated band. 
tb. Tongue-bars, vel. Velum. 

thymns gland. The thymus is one of those enigmatical 
ductless glands which are so eminently characteristic of 
the Vertebrate organisation, and are of the utmost phys- 
iological and pathological importance to the individual. 
In their structure and development they give clear indi- 
cations of having undergone an extensive change of 
function in the course of their evolution. 

The thymus, therefore, is presumably the derivative of 
an ancestral organ, which formerly possessed an active 
function as opposed to the apparently passive function 
which this gland, and others like it, exercise in the exist- 


ing Craniota. Amphioxus has hitherto been regarded 
as forming a marked exception among the Vertebrates 
in having no thymus, whereas one might reasonably have 
expected to find here the representative of the thymus 
in full activity. Although contrary to the prevailing 
impression, I would suggest that the thymus is repre- 
sented in Amphioxus by the very actively functional 

Do.. IN has shown that in the Selachian (shark) 
embryo the thymus arises by a series of distinct cell- 
proliferations from the epithelium of the dorsal wall of 

the successive gill-slits with 
the exception of the first, 
which is the spiracle (Fig. 
12). Sometimes these pro- 
liferations cause a small pro- 
jection downwards into the 
gill-slit, comparable to an 
incipient tongue-bar. Event- 



ually these separate thymus 
rudiments pass inwards and 
come together so as to form 
the definite thymus gland. 
Dohrn concluded from its 

Fig. 12. Horizontal section through mode of Origin that the 
the branchial reeion of an embryo of ,, i*. i r 

Scyllium canicula Tto show the rudiments thymilS resulted from the 

of the thymus. (After DOHRN.) metamorphosis and intro- 

sp. Spiracle, cav. Cavity (coslom) of 

branchial bar. I, II, III. First, second, Version of gill-filaments ; and 

and third gill-pouches, jug.v. Jugular . - f , , . . f 

vein. thy. Thymus rudiments. ln P C IaCt > L 

its morphological nature is 

probably correct. But the tongue-bars of Amphioxus, 
which correspond closely in position to the thymus rudi- 
ments in the Selachian embryo, and are, like the latter, 


essentially epithelial structures, are nothing else than gill- 
filaments or gill-lamellae. It appears, therefore, that we 
are justified in supposing that the tongue-bars of Amphi- 
oxus are the functionally active organs, of which the thymus 
of the higher forms is a metamorphosed derivative. 


Returning, then, to the consideration of the more inti- 
mate structure of the pharynx, --the endostyle has, been 
already mentioned as a ven- 
tral groove of the pharynx 
accompanying the latter 
throughout its whole length. 
A transverse section of it 
alone is shown in Fig. 13. 
It is composed of very high 
columnar cells arranged 
throughout in one layer, al- 
though the tenuity of the 
cells, whose nuclei are often 
placed at different levels, 

p-ives rise to the imnression Fig> '3- -Transverse section through 
S 1V endostyle of Amphioxus. (After LAN- 

of cells occurring in several KESTER. slightly altered.) 

e.a. Branchial artery with blood-clot, 
layers. The four groups of e .c. Endostylar ccelom. sk. Skeletal plate. 

gland-cells, placed symmet- 
rically two on either side of the median line, are the 
distinguishing feature of the endostyle. The cells are 
all ciliated, but those in the middle line bear a bunch of 
specially long cilia, which are of great importance in 
putting in motion the cord of mucus secreted by the 
glandular cells of the endostyle. Below the endostyle, there 
is a well-defined portion of the true body-cavity in which 
the branchial artery lies. This is the endostylar cozlom. 


Besides the rods in the gill-bars, there is a series of 
paired skeletal plates lying immediately below the endo- 
stylar epithelium (Fig. 13). These plates correspond in 
number to the primary gill-slits. Their shape and arrange- 
ment are shown in Fig. 14. They slightly overlap each 

other, and alternate 
with one another just 
as the primary gill-slits 
alternate. This alter- 
nation of paired struc- 
tures is of very general 
occurrence in Amphi- 
oxus, and affects almost 
every system of organs, 
--such as muscular, 

Fig. 14. Lower portions of skeletal rods of nerVOUS reproductive 
pharynx with three pairs of endostylar plates, 

seen from above. (After SPENGEL.) and branchial systems. 

The substance of the skeletal rods passes into j , , 

that of the endostylar plates (<?./). thus producing : maybe Stat( 

an arcade like the cover of a shoe (Spengel). creneral rule, tO which 
sy. Cross-bars (synapticula). 

there are some excep- 
tions, that with regard to the paired organs of Amphioxus, 
the organs of one side (e.g. myotomes, primary gill-slits, 
gonads, spinal nerves) do not lie opposite to their antimcres 
on the other side, but alternate with them. 

Branchial Bars. 

The structure of the branchial bars is shown in section 
in Fig. 15. Both kinds of bars, primary and secondary, 
have the same general appearance, being compressed and 
band-like, but the secondary bar is the smaller of the two. 

The chief point of difference between them is, that in 
the primary bar a portion of the ccelom is involved, which 
is absent in the secondary bar. In the case of the primary 



bar (Fig. 15 B}, commencing from the outside, that is to 
say, from the edge turned towards the atrial cavity, we 
have first a patch of columnar atria) epithelium, at the cor- 
ners of which some of the cells contain a quantity of the 
rich brown pigment which has been referred to above as 
being characteristic of the atrial epithelium generally. 

Fig. 15, A and B. Transverse sections through primary (B) and secondary 
(.4) gill-bars. (After BENHAM, slightly altered.) 

a.e. Atrial epithelium, b.e. Branchial epithelium, c. Ccelomic space of primary 
bar. sk. Skeletal rods. v. Ccelomic vessel of primary bar. v". External vessel 
of both bars, v'" . Internal vessel of both bars. 

N.B. Benham holds the space at the inner edge of the skeletal rod of 
tongue-bar for a blood-vessel. 

C. Isolated ciliated cells of the branchial epithelium. (After LANGERHANS.) 

Next comes a cavity which is a portion of the coelom, and 
is lined by the flat ccelomic epithelium. In fact, the dor- 
sal, or subchordal ccelom on each side (cf. Fig. 2) is put 
in connection with the endostylar coelom by a canalicular 
detachment of the ccelom which accompanies each primary 


bar of the pharynx. Wedged in between the ccelomic and 
atrial epithelia of the primary bar is a small blood-vessel, 
v. Internal to the coelomic space lies the skeletal rod, 
which in section has the shape of a triangle, at whose apex 
there is another blood-vessel, v". The sides and inner 
edge of the bar are composed of the ciliated pharyngeal 
epithelium. The cells of the latter are always arranged in 
a single layer, but at the sides of the gill-bars they are 
very long and thin, and the nuclei are crowded together at 
different layers so as to give the idea of a many-layered 
epithelium (Fig. 15 C). The cells of one side of the bar 
are in juxtaposition with those of the opposite side, except 
at a point near the internal edge of the bar, where a space 
occurs. In this space there is a third blood-vessel, v' n '. 6 

In the secondary bar, there is no vessel corresponding 
to the one marked v in the primary bar, and the vessel 
that corresponds with v" is entirely enclosed within the 
skeletal rod. 

The dorsal wall of the pharynx is closely appressed 
against the sheath of the notochord, and separates the two 
dorsal aortae from one another. It has here the form of a 
groove running parallel with and opposite to the endostyle. 
It is known as the hyperbranchial groove. (Cf. Fig. 2.) 
Two special tracts of ciliated epithelium form the sides of 
it, and pass downwards in front to join the anterior extrem- 
ity of the endostyle on each side. In front, where these 
tracts bend downwards with a crescentic curve, they are 
known as the peripharyngeal bands. (See Fig. n, ///./;.) 


The musculature of Amphioxus is composed almost 
entirely of striated muscle-fibres. Involuntary or smooth 
muscle-fibres are remarkable for their extreme tenuity, and 


in correlation with this condition is to be noted the absence 
of a distinct sympathetic nervous system. 

The striated muscles can be arranged in two groups : 
(i.) the parietal muscles constituting the myotomes or 
muscular segments of the body, and (ii.) the visceral 
muscles which arise independently of the myotomes and 
are not segmentally arranged. The smooth muscle-fibres, 
which occur on the walls of the alimentary canal and 
blood-vessels, may be grouped together as the splanchnic 

The parietal muscles are the great longitudinal muscles 
which make up the thick lateral walls of the body. In 
Amphioxus they form collectively the essential organ of 
locomotion. The portion of them lying next to the atrium 
on each side, and stretching from the notochord to the 
base of the myotome, is placed at an angle to the rest, and 
has a more vertical direction. (Cf. Fig. 2.) This has 
been described by SCHNEIDER as the rcctus abdomiuis. It 
probably co-operates with the muscles of the floor of the 
atrium to cause the contraction of the latter cavity for the 
purpose either of expelling water or reproductive elements 
through the atriopore. 

The visceral muscles consist of (a) the transverse muscles 
stretching across the floor of the atrium (cf. Fig. 2), (/3) 
muscles of the oral hood and cirri, (7) sphincter muscle of 
the velum (cf. Fig. 7), (>) anal sphincter. 

All the striated muscles of Amphioxus are composed 
of highly characteristic flat lamelliform plates, which can 
often be resolved into a great number of finer fibrils. In 
the longitudinal muscles of the adult, nuclei are very 
rarely met with, but in other places they are to be found ; 
as, for instance, in the fibres composing the velar sphincter 
(Fig. 1 6). 


This peculiar plate-like muscular tissue is found in 
connection with the lateral muscles only of the Cyclostome 
fishes. The muscle-fibres of the mouth 
and velum, as LANGERHANS pointed out, 
closely resemble those found in the walls 
of the heart of the higher Vertebrates. 
In transverse section the cut edges of 
the longitudinal muscle-plates are to be 
seen stretching across the myotome. 
(Cf. Figs. 2, 26.) 

The transverse or sub-atrial muscles are 
divided by a median longitudinal septum 
of connected tissue into right and left 
halves. They are further subdivided into 
a series of compartments by thin trans- 
verse septa. These compartments, how- 
ever, are not arranged segmentally, since 

Fig. 16. Isolated 

muscle-fibre of the they are more numerous than the myo- 
tomes. The muscle-plates of these mus- 

cles are placed edge on, so that they do 
not lie one over the other as the plates of the myotomes 
do, but one behind the other. They are attached to the 
septum at the base of the myotomes on the one hand, and 
to the median septum or raplic on the other, and also they 
are attached at numerous points to the connective-tissue 
sheath or fascia which covers them above and below. 
When they contract, therefore, the floor of the atrium is 
thrown into a number of characteristic pleats. (Cf. 
Fig. 2.) The individual muscle-plates of Amphioxus ap- 
pear universally to be devoid of a protecting sheath or 
sarcolemma. The sub-atrial muscles end at the atriopore, 
round which they form a sphincter muscle. 

The muscles of the oral hood, which serve for the erec- 



tion and supination of the buccal cirri, consist of two por- 
tions, an inner and an outer (Fig. 17). The outer one, 
by whose contraction the cirri are retracted in such a way 
that they come to lie across the entrance to the mouth, 
those of one side interlacing 
with those of the other so 
as to form a perfect barrier 
to the mouth, is a powerful 
muscle lying outside the 
bases of the cirri. The 
inner muscle, which appar- 
ently serves to erect the 
cirri, consists of distinct 
muscular tracts lying be- 

whose fibres interlace with those of the 

tween every tWO Consecutive velar sphincter (z/.#). /./. Inner muscle 
c j n -i (m. internus). 

The sphincter muscle of the velum has been already 
referred to. (Cf. Fig. 7.) A sphincter muscle of a simi- 
lar character also surrounds 
the anus. 

The septa which separate 
the myotomes from one 
another are composed of 
fibrous connective tissue. 
The fibres are imbedded in 

Fig. 17. Muscles of the oral hood. 

m.e. Outer muscle (m. externus) 













Fig. 18. -Diagram illustrating the ge l a tinOUS matrix, 
different layers of the integument. (After 

HATSCHEK.) salient feature in connexion 

/. Epidermis. 2. Outer layer of cutis 
(basement membrane of Hatschek and wlt " the entire Connective 

Sp H-T IK 3 ' Midd i le lay T" f cu l is whh tissue-system of Amphioxus 

radial fibres. 4. Inner layer of cutis. J 

5. Epithelial layer of cutis (limiting mem- is the great preponderance 


of the gelatinous element. 

It forms the bulk of the dorsal and ventral fin-rays, and 
of the cephalic and caudal integumentary expansions. 



The middle layer of the cutis below the epidermis (cf. 
Fig. 1 8) is composed mainly of this tissue with radial 
fibres superadded. In the metapleural folds it attains a 
greater development than in the rest of the integument. 
(Cf. Fig. 2.) It also constitutes the middle layer of the 
sheath of the notochord, but the fibres in this case run 
concentrically, and not radially.* The outermost layer 
of the cutis (Fig. 18) and the innermost layer of the 
sheath of the notochord are composed of a peculiar and 
very highly refringent and homogeneous tissue of the 
same order as that which forms the skeletal rods of the 
pharynx. 6 The layer of connective tissue which separates 
the myotomes from the body-cavity, and which springs 
out from the base of the notochordal sheath (Fig. 2), occu- 
pies the same position as the ribs of the higher Vertebrates. 


i. (p. 15.) Metapleural Folds. --\\\ the development of the 
paired fins of Selachians it was discovered, in 1876, by BALFOUR, 
that at a certain stage there appears along each side of the body 
"a thickened line of epiblast (i.e. ectoderm), which from the first 
exhibits two special developments." " These two special thick- 
enings are the rudiments of the paired fins, which thus arise as 
special developments of a continuous ridge on each side, precisely 
like the ridges of epiblast which form the rudiments of the 
unpaired fins." After giving more details, Balfour says, "The 
facts can only bear one interpretation, viz. that the limbs arc the 
remnants of contimums lateral fins."' 1 

Shortly afterwards (1877), but quite independently, JAMES K. 
THACHER was led by a comparative study of the adult skeleton of 

* In that portion of the sheath of the notochord which lies above the dor- 
sal groove of the pharynx thers is a special tract of connective-tissue fibres 
which run longitudinally. A similar tract can sometimes be observed in the 
dorsal portion of the sheath below the nerve-cord. (Schneider, Lankester, 

NOTES. 39 

Selachians and other fishes, to a belief in the homodynamy of 
median and paired fins, and he therefore concluded that the 
latter arose as differentiations from primitively continuous lateral 
fins just as the median fins are obviously differentiated from a 
continuous dorsal and ventral fin-fold. Thacher further suggested 
that the original continuous lateral fins were represented in Am- 
phioxus by the metapleural folds. He said, " As the dorsal and 
anal fins were specialisations of the median folds of Amphioxus, 
so the paired fins were specialisations of the two lateral folds 
(metapleural folds), which are supplementary to the median in 
completing the circuit of the body." 

It has recently been observed by Professor E. A. ANDREWS, that 
in a new species of Amphioxus from the Bahamas, the right meta- 
pleural fold is continued behind into the median ventral fin. 
Subsequently I found the same condition to obtain in a species 
of Amphioxus from the Torres Straits. 

From these observations, and from the fact that the right half 
of the oral hood (which apparently arises in continuity with the 
right metapleur ride infra} is continued in front into the 
cephalic expansion of the dorsal fin, it would appear that there 
is a great measure of truth in Thacher's suggestion, notwithstanding 
the fact that in the condition in which we find them in the exist- 
ing Amphioxus, the metapleural folds do not function as fins. 
Thacher's hypothesis has 
also been supported by 

2. (p. IO.) The aCCOm- Fi g I9 ._Diagram illustrating (by a dotted 

panying diagram (Fig. 19) line) the course of the food as it passes through 

.,, -,, ,, the pharynx and intestine of Amphioxus. (After 

will serve to illustrate the AN p REA ^ s \ 

actual course of the food The small diverticulum on the dorsal side of 

through the pharynx o f the oral hood represents the 

somewhat exaggerated. The arrows behind the 

AmphioxtlS, as recently caecum indicate the rotation to which the food is 
determined bv ANDREWS nere s "t>J ecte d by the action of the cilia of the 

intestinal epithelium. 
from observations on 

transparent specimens from the Bahamas. The food, enveloped 
in the mucous secretion of the endostyle, passes along the dor- 
sal groove of the pharynx (hyperpharyngeal groove) into the 


3. (p. 10.) On those occasions on which Amphioxus is not 
buried in the sand, but lies on the surface of the sand, occasions 
which frequently occur when it is kept in captivity, and especially 
after having been confined for a considerable length of time, it 
lies on one side, as mentioned in the text. The percentage of 
instances in which it lies on the right or left side has not been 
taken, and consequently it is not possible to say that it prefers lying 
on one side rather than on the other. Since the olfactory pit and 
the anus occur on the left side, it is conceivable that it prefers 
to lie on the right side. If this had been a definite habit, it 
would probably not have escaped the observation of Johannes 
Miiller. It is a fact which is too frequently overlooked, that the 
lying on one side is entirely incidental, and is emphatically not 
the result of adaptation to a peculiar mode of life, as it is in the 
case of the Pleuronectidae. 

4. (p. 1 1 .) Species and Distribution of Amphioxus. A useful 
synopsis of the genus Branchiostoma has recently been prepared by 
ANDREWS, as an appendix to his paper on the remarkable species 
which occurs at the Bahamas. In this species there is a long 
caudal process into which the notochord extends. It is an active 
swimmer. Gonadic pouches are only present on the right side, 
those on the left being suppressed. The latter is also true of 
Branchiostoma cultclhim. The peculiarities of the species from 
the Bahamas were such that Andrews deemed it necessary to form 
a new genus, Asymmetron. 

In the table of species on page 41 it will be noticed that the 
lengths of the different species are not in any proportion to the 
number of myotomes. 

Insufficiently described species occur off the coasts of Japan, 
Ceylon, and Fiji Islands. It is interesting to note that while in 
Europe, Amphioxus occurs as far north as Scandinavia, on the 
Atlantic coast of North America, Chesapeake Bay appears to be 
its northern limit, and it is therefore wholly unknown at the 
Marine Biological Station at Woods Holl. Attention may further 
be called to the simultaneous occurrence of two distinct species, 
B. ciiltelhim and B. belcheri, in the Torres Straits. B. cultellum 
is easily recognisable on account of the unusual height of its dorsal 








i. D. lanceolatum . 



Scandinavia, Heligoland, Eng- 

land, France, Mediterranean, 

and Chesapeake Bay. 

2. B. caribceum 



Brazil, Mouth of La Plata, 

Jamaica, Tampa Bay, Gulf of 

Mexico, Beaufort, N.C. 

3. B. cultellum . . 


2 5-35 

Thursday Island (Torres 

Straits), Moreton Bay (E. 


4. B. bassanum 


Bass Straits, Australia. 

5. B. belcheri . . . 



Borneo and Torres Straits 

(Prince of Wales Island). 

6. B. elongatum . . 




7. B. californiense 



San Diego, California. 

8. B. lucayanum . . 



Bimini and Nassau Harbour 

= Asymmetron lu- 


cayanum, Andrews 

5. (p. 22.) HUXLEY had recognised in 1874, in the light of 
Kowalevsky's work, that the atrial cavity of Amphioxus was lined 
by an epithelial layer derived from the ectoderm, but came to the 
conclusion that it was, by the very fact of its inversion within the 
body, converted into peritoneal epithelium. He applied the same 
interpretation to the opercular chamber of the Amphibian tadpole, 
and gave to a body-cavity of this character the general name of 
epicene. ROLPH'S merit consisted in distinguishing clearly between 
atrial epithelium and peritoneal epithelium, and hence between 
atrial cavity and true body-cavity. 

6. (p. 38.) There is a great deal of difference of opinion as to 
the exact nature of that dense refringent tissue which forms the 
outer layer of the cutis and the skeletal rods of the gill-bars. 
LANKESTER regarded them both as the products of connective 
tissue-cells, hence belonging to the mesoderm, while HATSCHEK 
and SPENGEL looked upon the outer layer of the cutis as the 
product of the ectoderm, of the nature of a basement membrane. 
SPENGEL again has advocated the view that the skeletal rods of the 


pharynx are special developments of the basement membrane, 
which separates the two opposed epithelial layers of each gill-bar 
from one another. (Cf. Fig. 15.) More recently, BENHAM has 
described nuclei in the latter membrane, thus showing it to be a 
sheet of connective tissue. In this case the substance of the 
skeletal rods should be regarded as a variety of connective tissue. 

A further difference of opinion prevails as to the nature of the 
space which traverses the skeletal rod of the tongue-bar. LAN- 
KESTER supposed it to be a diverticulum of the coelom. SPEXGEL 
and BOVERI interpreted it as a blood-vessel ; and, finally, BENHAM 
thinks that it is both, inasmuch as he conceives there to be a blood- 
vessel contained in a ccelomic space. It should be added that 
these finer details are extremely difficult to determine. 

7. (p. 21.) Lateral Line. Since the lateral line constitutes 
one of the most characteristic and constant features in the organi- 
sation of fishes, its absence in Amphioxus has always been one of 
the most serious difficulties in the way of a conception of this 
animal as, in any sense, an ancestral form. It need hardly be 
pointed out that from whatever point of view we regard Amphi- 
oxus, it must necessarily have become specialised and modified 
along its own particular line of evolution, and cannot, as it stands, 
be taken as a direct ancestral form, but rather as a more or less 
close relative of, or an exceedingly ancient offshoot from, the 
actual ancestor of the Vertebrates. The modifications which it 
has undergone will, as in every other case, have resulted in more 
or less extensive changes both in the function and structure of dif- 
ferent parts. Thus, while the metapleural folds are very probably 
the homologues of the primitive continuous lateral fin-folds, yet 
in their actual form and function they may or may not represent 
the primordial condition of these folds. Certain peculiar features 
in connexion with the origin and innervation of the metapleural 
folds of Amphioxus have led me to form a conception as to the 
origin of the lateral line sense-organs which may perhaps have 
some value as a working hypothesis. 

In those primitive fishes which possessed the continuous lateral 
fin-folds, it is very clear that the latter could not have performed 
a locomotor function, but they must have served primarily as 
balancers. Without going into the difficult question as to how 

NO TES. 43 

such structures could have arisen de novo, we may at least attempt 
to appreciate the necessity for their existence. 

There is one difference between the general form of the body 
in Invertebrates and Vertebrates respectively which seems to be 
of fundamental importance, but which has not been sufficiently 
emphasised. As a general rule, in the Invertebrates, the body 
is not bilaterally compressed, but, on the contrary, is either cylin- 
drical, sub-cylindrical, or flattened dorso-ventrally. Obvious ex- 
ceptions to this rule are presented by the Lamellibranchiate 
Molluscs and by many Arthropods ; but these exceptions are 
readily intelligible as secondary modifications. 

On the other hand, in the more primitive Vertebrates (i.e. 
fishes), the bilateral compression of the body is one of the car- 
dinal features of the external form. To this fundamental rule 
there are of course exceptions afforded, for example, by the 
skates ; but it is a self-evident fact that these again have arisen 
by secondary modification from originally bilaterally compressed 
forms. With the evolution of the pentadactyle appendages and 
the assumption of a terrestrial existence, the shape of the body 
in the higher Vertebrates has undergone such changes that the 
primitive bilateral compression of the body is, as a rule, only 
present at some period of the embryonic development. 

Amphioxus exhibits the characteristic vertebrate bilateral com- 
pression of the body in a very typical manner ; while Balano- 
glossus shows invertebrate affinities in regard to the shape of the 
body, which is sub-cylindrical. 

The bilateral compression of the primitive vertebrate body did 
not arise in itself as a special adaptation to a particular mode of 
life ; but rather in correlation with other characters of the organi- 
sation. The development of the dorsal medullary tube and the 
notochord above the digestive tube and the concentration of the 
myotomes would necessarily lead to a bilaterally compressed form 
of body. We see this not only in fishes, but in the course of the 
development of all Vertebrates. 

It is obvious that such a shape of the body is highly unfavourable 
for the maintenance of the equilibrium except with the assistance 
of some special mechanical and sensory apparatus. 

Now in Amphioxus, the metapleural folds, whatever their exact 


function may be, do not serve in any way as balancers ; and, as 
mentioned in the text, Amphioxus has no means of maintaining 
its equilibrium when not actually swimming. 

We will therefore keep in mind more especially those Palaeo- 
zoic fishes which presumably possessed continuous lateral fin-folds 
serving as balancers. The nearest known fossil relatives of these 
fishes appear to be the Cladoselachidce (see BASHFORD DEAN. 
Contributions to the Morphology of Cladoselache ( Cladodus) . 
Jour. Morph. IX. 1894. pp. 87-112. Also A. SMITH WOODWARD. 
The Evolution of Fins. Natural Science, I. 1892. pp. 28-35). 

The lateral fin-folds may be spoken of as mechanical balancers, 
and to render them efficient organs, there must be a sensory appa- 
ratus in connexion with them. The suggestion lies near that the 
ectoderm which took part in the formation of the lateral fin-folds 
also produced the sense-organs of the lateral line. 

The lateral line, through its capacity for receiving impressions 
of wave-movements, etc., would thus serve as the agent in the 
co-ordination of such muscular activities as are necessary to the 
maintenance of the equilibrium. 

Having been once established, no special difficulty is presented 
by the fact that the lateral line has spread over the head-region. 
Moreover, it may be taken as a well-established morphological 
fact that the auditory organ (internal ear) became evolved as a 
specialisation of part of the lateral line in the cephalic region, and 
that it therefore belongs to the same category as the less elaborate 
sense-organs of the remainder of the lateral line. 

As is well known, the internal ear has two functions, audition 
and equilibration. It must be supposed that, at its first origin, 
the whole lateral line served in a general way the function of 
equilibration, and that this function eventually became chiefly 
localised in the semicircular canals of the ear, the remainder of 
the lateral line perhaps undergoing a slight change or limitation 
of function. 

It seems certain that at first the sense-organs of the lateral line 
must have been innervated by spinal nerves. This follows both 
from a priori considerations and also from the condition in Amphi- 
oxus, where the ectoderm of the metapleural folds is innervated 
by the Rami cutanei ventrales of the dorsal spinal nerves. Under 

NOTES. 45 

these circumstances it is necessary to suppose with EISIG that the 
lateral line nerve (Ramus lateralis vagi) arose as a collector. 

The removal of the lateral line from the immediate neighbour- 
hood of the paired fins in existing fishes is easily intelligible on 
the ground that the fins have become discontinuous and elaborated 
into effective locomotor organs. 

It is not impossible that the lateral line nerve (ft. late rails vagi) 
is homodynamous with the remarkable Ramies cutaneits quiuti 
(R. recurrent trigemini et facialis or Nervus lateralis frige mini, 
STANNIUS) of Teleosteans, which runs to the base of all the fins, 
paired as well as unpaired ; just as the paired fins themselves are 
known to be homodynamous with the median fins. In this case 
the R. cutaneits quinti would be of primitive significance, notwith- 
standing the fact that it is absent in Selachians ; and it would be 
another of those features of organisation in the possession of which 
Teleosteans exhibit more primitive relations than do the existing 
Selachians. (Compare the functional pronephros of Teleosteans 
and the entirely rudimentary pronephros of Selachians.) 

The above suggestion that the lateral line arose in the first 
instance as a sensory equilibrating apparatus in conjunction with 
the mechanical equilibrating apparatus effected by the continuous 
lateral fin-folds, will of course meet with numberless difficulties 
when it is attempted to carry it out in detail. As in some other 
respects, so here, a great difficulty is presented by the Cyclo- 
stoines. It may, however, be pointed out that if the various con- 
clusions which have been drawn with regard to the morphology of 
Amphioxus are correct, it must be assumed that the Cyclostomes 
have entirely lost the lateral fin-folds and that the sense-organs 
of the lateral line have secondarily become diffused in their dis- 
tribution over the body. The latter conclusion is also indicated, 
firstly, by the fact that there is a fairly well developed internal 
ear in the Cyclostomes which, as noted above, must have been 
differentiated from a primitive lateral line ; and secondly, by the 
fact that although the sense-organs are scattered, there is never- 
theless (at least in Petromyzon) a definite lateral line nerve. 



INTERNAL ANATOMY (continued). 

IN the preceding chapter we have seen how Amphioxus, 
while possessing the general facies of a fish, and the 
primary essential attributes of a Vertebrate, is nevertheless 
destitute of many of the most obvious structural features 
which we usually associate with our conception of a fish. 
Thus it has no skull, or, in other words, it is Acmniate 
(Haeckel). It has no jaws, and is therefore a Cyclostome, 
as opposed to a Gnathostome. Finally, it has no paired 
sense-organs and no paired muscular fins. Its eye-spot 
is median, like that of a Cyclopean monster. There is no 
trace of an auditory organ of any kind, while the single 
so-called olfactory pit, abutting on the anterior end of the 
nerve-tube, has been regarded as an indication of a mono- 
rhinic condition preceding the amphirhinic, i.e. with paired 

Vascular System. 

Now, in turning our attention to the vascular system, we 
shall find that Amphioxus has no heart. In any ani- 
mal with a comparatively well-developed vascular system, 
the presence of a heart might be regarded as a sine qua 
non. This, however, is by no means always the case ; and 
although, among the Invertebrates, the extensive groups 

4 6 


of the Arthropoda (Insects and Crustacea) and the 
Mollusca are characterised by the possession of a definite 
muscular heart, yet in the various groups of worms there 
are many which possess a very elaborate vascular system, 
while not one of them possesses a heart. In fact, in the 
last-mentioned forms, the place of a heart is taken, func- 
tionally, by contractile blood-vessels. And this is the case 
with Amphioxus. Among the Vertebrates, including 
the Ascidians, it forms the unique instance in which such 
an acardiac condition of the vascular system is met with. 

Lying below the pharynx in the endostylar ccelom, there 
is a blood-vessel known as the bmncJiial artery, which con- 
tracts more or less rhythmically, and corresponds in its 
position and relations to the heart and truncus arteriosus 
of the higher forms. 

Fig. 20. Diagram illustrating the chief parts of the vascular system of 
Amphioxus. (Constructed after J. MULLER and SCHNEIDER.) 

The arrows indicate the direction of flow of the blood, ch. Notochord. hep. 
Hepatic ccecum. af. Afferent branchial vessels (vascular bulbils of J. Miiller) 
entering the primary bars from br.a, the branchial artery ; the efferent branchial 
vessels are seen emerging from the tops of both primary and secondary bars and 
running into d.a, the dorsal aorta. From the dorsal aorta, the blood enters the 
capillaries over the wall of the intestine (indicated by the dark reticular shading), 
and finally reaches s.i.v, the sub-intestinal vein. The latter carries the blood to the 
base of the hepatic ccecum, over which it passes into another system of capillaries 
(not indicated), and is then collected into A.v, the hepatic vein, which passes back- 
wards and curves round into the branchial artery. 

From this branchial artery, lateral branches running up 
into the primary bars of the pharynx are given off on both 
sides alternately. (Cf. Fig. 20.) There appears to be no 

4 8 


direct communication between the vessels of the tongue- 
bars and the branchial artery. 

At the base of the primary bars the lateral offshoots of 
the branchial artery are found to be enlarged to form 
vascular bulbils, which are also contractile. Furthermore, 
at this point they divide into three branches of smaller 
calibre, which constitute the vessels of the primary bar. 



Fig. 21. Diagram of a section through the pharynx involving a primary bar 
(to the left), and a tongue-bar (to the right), to illustrate the circulation in the 
branchial bars. (After SPENGEL.) 

br.a. Branchial artery, c. Ccelom ; outside of which is the atrial epithelium. 
c.v. Coelomic vessel of primary bar. e. Endostyle. e.c. Endostylar ccelom. e.v. 
External vessel, i.v. Internal vessel. La. Left aorta, r.a. Right aorta. /. Cavity 
of pharynx, t.b. Tongue-bar. 

(Cf. Fig. 15.) One of these branches, as we have seen, 
runs up between the coelomic and atrial epithelium, and 
may be called after BOVERI the ccelomic vessel, of the primary 
bar (Figs. 15 and 21). Another lies at the inner edge of 
the skeletal rod, and is the so-called external vessel, while 
a third lies immediately below the inner pharyngeal epi- 
thelium of the bar, and forms the internal vessel. 


The two last-named vessels only are represented in the 
tongue-bars, and differ in their arrangement in the latter 


in so far as the external vessel is enclosed within the 
skeletal rod. 

The blood which circulates in the tongue-bars flows into 
them, not from the branchial artery, but from the primary 
bars through the cross-bars of the pharynx. The vessels 
of each gill-bar unite above into a single efferent vessel, 
which conducts the blood into the dorsal aorta of either 
side. So that while efferent vessels issue alike from both 
primary and tongue-bars, the afferent vessels, which lead 
the blood directly from the branchial artery into the gill- 
bars, are confined to the primary bars (Fig. 20). The 
blood, having been oxygenated during its passage through 
the gill-bars, past which a constantly renewed stream of 
water is kept flowing, enters the dorsal aorta, and is then 
carried backwards to the region of the intestine. The 
two halves of the dorsal aorta, which we have already 
noted on either side of the hyperpharyngeal groove, be- 
come united into a common trunk behind the pharynx, so 
that in the region of the intestine there is a single dorsal 
aorta (cf. Fig. 28), from which lateral branches are given 
off to the wall of the intestine. These then break up into 
capillaries, which anastomose freely together, and so form 
a perfect vascular network round the intestine. Finally, 
the blood emerges from this capillary system into a large 
vein lying below the digestive canal, the sub-intestinal vein. 
Here it flows in a forward direction until it reaches the 
base of the hepatic ccecum. At this point the vein appears 
to stop short, but in reality breaks up into another system 
of capillaries surrounding the liver. 1 From these again 
the blood is collected into the large multiple hepatic vein 
above the ccecum. Here it flows backwards as far as 



the angle formed by the ccecum with the alimentary canal, 
where the vein bends sharply round into the branchial 
artery, and so the cycle is completed (Fig. 20). According 
to JOHANNES MULLER, the time required for one complete 
circulation of the blood in Amphioxus is one minute, and 
in this time any given droplet of blood will have traversed 

the whole body. Con- 
trary to what takes place 
in the higher Verte- 
brates, a single contrac- 
tion of the heart (i.e. 
branchial artery) in 
r.a Amphioxus suffices for 
a complete circulatory 
cycle. 2 

The right and left 
dorsal aortas differ from 
one another in respect 
to the behaviour of their 
anterior cephalic termi- 
nations. At the front 

Fig. 22. Transverse section through re- rut. 

gion of velum to show difference in behaviour end Of the pharynx, the 


ch. Notochord. La. Left aorta, m. Meta- into a wide Vascular 6X- 
pleur. n. Spinal cord. r.a. Right aorta, t.m. . . 

Transverse muscles ; the septum (raphe) which pailSlOn which f 

divides these muscles into two halves is no ypJjjjY) QJ-J the ri"ht side 
longer median, but shifted towards the right 

side in consequence of the fact, discovered by (FigS. 3 and 22, I'.a.). 
VAN WlJHE, that the right transverse muscles . T\T 11 u 

dwindle out and end in this region, while the Johannes Muller, Wh 

left transverse muscles are continued into the j-gj- fjo-ured this StrUC- 
outer muscle of the oral hood. v. Velum. 

ture, took it for the an- 

teriormost aortic arch connecting the branchial artery 
directly with the dorsal aorta. 

However, according to the recent researches of Professor 


J. W. VAN WIJHE, it would appear that this so-called aortic 
arch does not communicate with the branchial artery, but 
ends blindly below in the neighbourhood of the right meta- 
pleur. Dorsally, the aorta from which this lateral arch-like 
outgrowth occurs, is continued forwards (not as a simple 
vessel, but as a complex of vessels) as far as a peculiar 
sense-organ known as the groove of Hatschek, after its 
discoverer. This groove lies in the roof of the oral hood 
to the right of the notochord, and is derived from the 
pr&oral pit of the larva (see below). (Cf. Fig. 76.) 

In front of the sense-organ this dilated continuation of 
the right aorta communicates beneath the notochord by 
means of a transverse vascular commissure with the left 
aorta, which retains its small calibre and simple character 
throughout. From the vascular complex of the right 
aorta arise the vessels which supply the buccal cirri. 

Hitherto we have only spoken of those blood-vessels 
which are related to some part or other of the alimentary 
canal. In point of fact the parietal or somatic vessels of 
Amphioxus, if present at all, must have a very subordi- 
nate physiological significance. Their place is taken by 
lymph-spaces, of which there are a great number in various 
parts of the body. Such are the dorsal and ventral fin- 
chambers, the spaces in the metapleural folds, spaces at 
the apices of the myotomes and in connexion with the 
dorsal nerve-roots, etc. (Cf. Fig. 2.) 3 

The vascular system of Amphioxus presents several 
features of great interest from a phylogenetic or evolu- 
tionary point of view. 

We have seen that the heart is in no way differentiated 
from the branchial artery and is therefore a simple tubular 
vessel. This is the primary condition of the heart in the 
embryos of all the craniate Vertebrates. In the latter, as 


the embryonic development proceeds, this simple tubular 
heart widens out, acquires a series of constrictions, and 
undergoes a remarkable flexure known as the signioid 
flexure. Two stages in the formation of the sigmoid 
flexure of the heart of the chick-embryo are shown in 
Figs. 23 and 24. At a somewhat earlier stage than 



Figs. 23 and 24. Anterior portions of chick-embryos of the 38th and 48th 
hour of incubation, seen from below, to illustrate formation of heart. (After 

ao. Right and left aortse. au. Auditory involution, c. 2 . Ventricular por'i -n of 
heart. c 3 . Auricular portion of heart, e. Eye. h. Heart, op. Priman ^tic 
vesicle, p.f.b. Primary fore-brain, p.m.b. Primary mid-brain, p.h.b. Pr. 
hind-brain, t.a. Truncus arteriosus. v.a. Vitelline arteries. v,v. Vitelline veni. 
i, 2,3, Transitory gill-slits. 

that represented in Fig. 23 the heart was pt ! ",?ctly 
straight. In this figure it is still a simple dilated tube, 
but no longer straight. It has become bent outwards 
into a U-shape. At the stage of Fig. 24 well-marked 
constrictions (the indications of the later division into 
auricle and ventricle, etc.) have appeared in the heart, and 
the simple U-shaped flexure of the latter has become 


complicated by the occurrence of a further flexure in a 
different direction, in consequence of which the hinder 
limb of the U has been raised, so to speak, to nearly the 
same plane as the anterior limb. The shape of the heart 
at this stage bears a characteristic resemblance to the 
Greek letter sigma. The permanent condition of the 
heart in Amphioxus therefore corresponds to an early 
stage of its development in the higher Vertebrates. 

Again, in the craniate embryo the dorsal aorta arises as 
a pair of vessels on either side of the notochord, which 
later fuse together into one median dorsal vessel. (Cf. 
Fig. 24.) In Amphioxus, throughout a great portion of 
its extent, namely, in the region of the pharynx, the two 
halves of the dorsal aorta remain permanently separated 
from one another by the dorsal groove of the pharynx. 
(Cf. Figs. 2 and 28.) 

One of the most striking peculiarities of the vascular 
system of Amphioxus is the presence of the sub-intestinal 
vein, in its capacity as the main venous trunk of the body. 
It collects the blood from the capillaries of the intestinal 
wall, and conducts it to the base of the liver, where it again 
breaks up into capillaries.* It acts, therefore, physiologi- 
cally, as a portal vein, while morphologically it is the 
sub-intestinal vein. Curiously enough, it is much larger in 
its posterior than in its anterior moiety, and in transverse 
sections through the hinder region of the intestine there 
appear to be several separate vessels lying side by side, 
sometimes as many as six. These, however, if traced 
backwards or forwards, are found to anastomose with one 

* In the larva of Amphioxus the sub-intestinal vein and branchial artery 
form one continuous blood-vessel. Later, when the hepatic ccecum (liver) 
grows out from the ventral wall of the alimentary canal, an interruption occurs 
in the continuity of the vessel, through the insertion of a capillary portal system 
in its course. 




Fig. 25. View of portion 
of sub-intestinalvein of Amphi- 
oxus, to show its fenestrated 
character in the posterior re- 
gion. (After SCHNEIDER.) 

a. Anterior, p. Posterior. 

another, as shown in Fig. 25, and 
so there is produced a fenestrated 
structure in the vein. The hepatic 
vein has a similar fenestrated char- 
acter, and this was what was meant 
by speaking of it above as being 

The sub-intestinal vein reappears 
in the embryos of all the higher 
fishes and Amphibia, where it breaks 
up into capillaries in the liver. In 
these forms, however, it does not 
persist long as the main venous 
trunk, but becomes replaced almost 
entirely by the development of two 
large veins, which arise on either 
side of the dorsal aorta. These are 
the so-called cardinal veins. The 
sub-intestinal vein mostly disappears 
after the formation of the cardinal 
veins, but persists as a second-class 
vessel in the lampreys and in some 
sharks, lying, in the latter, in the 
spiral valve of the intestine.* More- 
over, its posterior portion, which 
lies in the tail, persists as the caudal 

* The sub-intestinal vein is also persistent in 
the following Urodele Amphibia Salaman- 
dra, Triton, and Pleurodeles. (See F. HOCH- 
STETTER. Beitrage zur vergleichenden Anatomie 
tind Entwicklungsgeschichte des Venensystems 
der Amphibien und Fische. Morph. Jahrb. 
XIII. 1888. pp. 119-172.) 


The same vessel, therefore, which constitutes the main 
venous trunk of the adult Amphioxus performs the same 
function in the embryos of the higher fishes. We can thus 
deduce a good deal of evidence from a consideration of the 
vascular system alone, pointing to the primitive and ances- 
tral character of Amphioxus. 

If we compare broadly the vascular system of Amphioxus 
with that of a segmented worm like the common earth- 
worm, we are at once confronted with certain obvious 
superficial resemblances. Here, as in Amphioxus, the 
vascular system comprises two main longitudinal trunks, 
one lying above the intestine and the other below it, and 
furthermore, they are connected together at intervals by 
circular vessels which form complete rings round the 
alimentary canal in the same way as do the vessels which 
pass through the pharyngeal bars of Amphioxus. 

It is only when we come to enquire into the direction 
of flow of the blood in the two cases that we meet with a 
striking contrast between them. Whereas in Amphioxus 
the blood flows in the dorsal aorta from before backwards 
(see Fig. 20), and in the sub-intestinal vein together with 
the branchial artery, from behind forwards, in the worm, on 
the contrary, these directions are reversed, and the blood 
flows from behind forwards in the dorsal vessel, and from 
before backwards in the ventral vessel. 

The Excretory System. 

The excretory function is so intimately bound up with 
the circulation that a description of the organs which 
serve this function follows naturally after the consideration 
of the vascular system. The apparent absence of definite 
excretory organs in Amphioxus was for a long time one of 
the greatest difficulties in the way of a correct appreciation 


of the peculiarities of its organisation. Thanks, however, 

to recent researches, it is now known to possess such 

organs in luxuriant abundance. 

From first to last several entirely different structures 

have been credited with a renal function. JOHANNES 

MULLER first discovered 
certain glandular epithe- 
lial tracts in the floor of 

"" cc the atrial chamber in its 

hinder portion. These 

cellular thickenings are. 
distinguished by their 
high cylindrical cells from 
the flattened atrial epi- 
thelium which surrounds 
them. (Cf. Figs. 1 1 and 
26.) Johannes Muller sug- 
gested that these groups 
of cells might be renal 
organs. His observation, 
however, failed to find 
general acceptance among 
morphologists for ;'bout 
thirty-five years, when, 

' m l ^7^> W. R 



Fig. 26. Transverse section through 
post-pharyngeal region of young individual, p AUL LANGERHANS, WOrk- 
to show groups of renal cells in floor of 
atrium. (After LANKESTER and WiLLEY.) ing independently, fully 

ao. Aorta, at. Atrium, b.c. Bodv-cavity ., , . 

(ccelom). c.c. Central canal of nerve-cord Confirmed hlS ECCOUH 
(.c). d.f.c. Fin-cavity. i.m. Intercoelic am j accep ted his inter- 
membrane. /.;. and r.m. Left and right 

metapleural folds, r.p. One of J. Muller's pretation of the bodies as 
renal papillae, s.i.v. Sub-intestinal vein. , , u 

renal organs, at the same 

time adding a careful histological description of them 
(Fig. 27). 



The individual groups of cells have an elongated and 
more or less ovoid shape with the long axis parallel to the 
long axis of the body. According to Langerhans their 
surface is ciliated. Two kinds of cells enter into their 
composition ; namely, large clear dilated cells, which are 
separated from one another by fine fibre-like cells of 
extreme tenuity (Fig. 27). In the latter the nucleus of 
each cell is placed near the free end of the cell, while in 
the former it lies nearer the 
base of the cell. Langerhans 
found highly refringent con- 
cretions in the dilated cells 
which he took for excretory 
products. That these cells 
have a capacity for excreting 
waste matters has more re- 
cently been shown experiment- 
ally by F. E. WEISS. The atrial 
epithelium on the pharyngeal Fig. 27. -isolated ceils from renal 

papilla ; the large cells contain con- 
bai'S has a similar Character cretions indicated by the black bodies. 

r . (After LANGERHANS.) 

to that forming these curious 

renal papillae on the floor of the atrium. The distribution 
of these papillae in the vicinity of the atriopore is very 
irregular and variable and without any regard to a sym- 
metrical disposition. Although they are undoubtedly to 
be regarded as a species of renal organ, yet they could 
not be compared to any portion of the excretory system 
of the higher Vertebrates. 

Another structure, or pair of structures, which has been 
considered to belong to the category of renal organs must 
next be referred to. 

This consists of two funnel-shaped diverticula of the 
atrial cavity lying in the dorsal (subchordal) ccelom in the 


region of the twenty-seventh myotome, where the pharynx 
ends and the intestine begins. They were discovered in 
1875 by LANKESTER, who called them the atrio-ccelomic or 
brown funnels, on account of the rich accumulation of 
brown pigment in their walls. We have already referred 
to this brown pigment as occurring very generally in the 
atrial epithelium. The brown funnels have the shape of an 


Fig. 28. Plastic diagram illustrating the positions and relations of the atrio- 
coelomic funnels. A rod is passed through the peri-enteric coslom into the sub- 
chordal (suprapharyngeal) coelom. (After LANKESTER.) 

ao. Dorsal aortas, at. Atrial cavity, b.f. Atrio-coslomic funnels, go. Gonads. 
l.d. Ligamentum denticulatum (pharyngo-pleural folds, Lankester). l.m. and 
r.m. Left and right metapleural folds. my. Muscles. pfi. Roof of pharynx. 
u. Point of union of the right and left aorta? into the median aorta. 

elongated cone, the apex of which is directed forwards. 
At the wide end each funnel opens into the atrial cavity, 
while at the narrow end it is possible, but not certain, that 
an opening exists into the dorsal coelom (Fig. 28). The 
funnels are adherent throughout their entire length to the 
roof of the dorsal ccelom. 4 


In 1889 WEISS undertook the task of determining ex- 
perimentally whether Johannes MUller's renal papilla and 
Lankester's brown funnels really served an excretory 
function. The method of research consisted in feeding 
full-grown individuals with various colouring matters held 
in solution or in suspension in sea-water. For instance, 
carmine suspended in sea-water would be carried into the 
digestive canal and then absorbed through the intestinal 
epithelium into the capillaries surrounding the intestine. 
It would thus get into the vascular system, and also by 
some means into some of the lymph spaces, and finally 
would be excreted by the cells of the renal papillae or by 
whatever other structure, or set of structures, might 
possess the renal function. In fact, Weiss found that the 
so-called renal papillae did actually excrete a quantity of 
the carmine with which the animals had been fed, and, 
further, that a similar excretion of carmine occurred at 
other points of the atrial epithelium. The atrial epi- 
thelium, as a whole, probably has more or less the power 
of excreting waste products which have found their way 
into the vascular and lymphatic systems. 

But above all, Weiss discovered a very active excretion 
of carmine in certain small tubules which he found lying 
in the dorsal ccelom applied against the most dorsal por- 
tion of the double-layered membrane (ligamenturn denti- 
culatum) which separates the ccelom from the atrial cavity 
(Fig. 29). There is one of these tubules to each primary 
gill-cleft of the pharynx. At the top of each tongue-bar 
Weiss made out an opening of the tubule into the atrial 
cavity, but he did not succeed in finding any openings into 
the dorsal coelom. After the operation of feeding with 
carmine was completed, at the close of a week or fortnight, 
and time had been allowed for its absorption and subse- 



quent excretion, the epithelium lining the walls of these 
tubules was found to be full of carmine granules. 

At about the same time at which Weiss was pursuing 
his studies on Amphioxus THEODOR BOVERI, having been 
led by independent a priori considerations, largely induced 
by the work of RUCKERT on the development of the ex- 
cretory system of Selachians, to suspect the occurrence 


Fig. 29. Portion of transverse section through the pharynx of Amphioxus, 
to show position of excretory tubule. (After WEISS.) 

ao. Left aorta, at. Atrial cavity, at.e. Atrial epithelium, c. Coelom. ch. Noto- 
chord. z.z. Interccelic membrane, l.d. Ligamentum denticulatum. nph. Excre- 
tory tubule, p.b. Primary bar. ph.e. Epithelium of hyperpharyngeal groove. 
ph.f. Pharyngo-pleural fold, s.c/i. Sheath of notochord. t.b. Tongue-bar. 

of excretory tubules in Amphioxus comparable to those 
found in the embryos of the higher Vertebrates, instituted 
a search for them and discovered them independently in 
the most brilliant manner. 

Boveri carried his investigation to a high pitch of per- 
fection, and has published an account of these tubules, 
which in point of clearness and completeness leaves nothing 



to be desired. The accompanying figures, taken from 
Boveri's finely illustrated memoir, show the appearance 
and topographical relations of the excretory tubules. 

A tubule as seen in the living condition is shown in 
Fig. 30. It is a curved tube consisting mainly of two 

Fig. 30. An excretory tubule of the left side, with the neighbouring portion 
of the pharyngeal wall, as seen in the living condition. The round bodies in ihe 
wall of the tubule represent carmine granules. Highly magnified. (After BOVERI.) 

limbs, bent approximately at right angles to one another, 
and lying over against the dorso-lateral wall of the phar- 
ynx. (Cf. Fig. 29.) The anterior limb is directed verti- 
cally, and the posterior longitudinally. The former opens 
by a relatively wide and forwardly directed opening into 


the dorsal coelom. The posterior end of the tube also 
opens into the coelom, and between these two terminal 
openings there is a variable number of other c&louiic 
openings, or funnels, as they are called, situated on the 
dorsal side of the tubule, and opposite to that side which 
carries the opening into the atrial chamber. The ccelomic 
funnels are placed at the ends of short upstanding projec- 
tions from the main body of the tubule. On the ventral 
side of the tubule, opposite in each case to a tongue-bar of 
the pharynx, occurs the single opening into the atrial cav- 
ity. The epithelium lining the tubule consists of cubical 
ciliated cells. There is a thick bunch of cilia in connec- 
tion with the atrial opening of the tubule. The curious 
thread-like structures, carrying a round knob at their dis- 
tal extremities, which radiate out from the ccelomic open- 
ings, are specially modified cells belonging to the ccelomic 
epithelium, which are probably concerned in promoting 
the excretory activity of the tubule, and are called by 
Boveri, tJiread-cclls (Fadenzellen). 

The vascular supply and exact location of the nephridial 
tubules (each tubule representing a nephridium, according 
to Lankester's nomenclature) are shown in Fig. 31. The 
figure represents a piece of the upper wall of the pharynx, 
cut out in such a way as to expose the inner wall of the 
dorsal coelom. The cross is placed at the cut edge of the 
double-layered membrane which separates the dorsal coelom 
from the atrial cavity. This cut edge can be traced from 
side to side of the figure. The membrane is seen to be 
continued down each primary gill-bar, in company with the 
extension of the ccelom, which runs down the primary bars 
into the endostylar ccelom as described above. On the 
other hand, the .membrane skips over the tongue-bars, so 
that the atrial cavity is prolonged dorsalwards into a deep 


bay, corresponding to each tongue-bar. (Cf. Fig. 29.) 
This is what produces the sinuous, or notched, appearance 
to the membrane in question, and led Johannes Miiller to 
speak of it as the ligamentum denticulatum, (Cf. Fig. 28.) 
The external or atrial opening of the tubule lies against 
the tongue-bar at the head of this bay-like extension of the 
atrial cavity (Fig. 31 on the right). 

The vascular supply of the tubules is effected in each 
case by the co-operation of two blood-vessels ; namely, the 

Fig. 31. Plastic figure illustrating the blood-supply (glomeruli) of the excre- 
tory tubules. On the right, the drawing is taken at a deeper level, to show the 
atrial opening of the tubule over against a tongue-bar. (After BOVERI.) 

>J<. Cut edge of ligamentum denticulatum. c.v. Coelomic vessel of primary bar. 
e.v. External vessel, i.v. Internal vessel, d.a. Left dorsal aorta. 

ca'lomic vessel of the primary bar (cf. Figs. 15 and 21) and 
the external vessel of the secondary, or tongue-bar. As 
soon as the coelomic vessel of a primary bar arrives at the 
level of a tubule, it gives off a number of branches, which 
not only anastomose among themselves, but become united 
with a similar series of anastomosing vessels which origi- 
nate from the external vessel of the next-following tongue- 


bar. In this way, a complicated plexus of blood-vessels is 
formed around and about the tubule. This vascular plexus 
is known as a glomcruhis. 

The blood charged with whatever waste matters it may 
have gathered up in its course through the body arrives 
eventually at the glomeruli, where it is considerably 
delayed on account of the vascular plexus through which 
it has to pass before reaching the dorsal aorta. During 
this delay, it is exposed to the glandular excretory action 
of the tubules, by which the waste products are extracted 
from the blood by osmotic action. From the glomerulus 
the blood is conducted by two efferent vessels, corre- 
sponding respectively to the primary and tongue-bars, 
into the dorsal aorta. The communication between two 
neighbouring glomeruli, as shown in Fig. 31, is, according 
to Boveri, the exception and not the rule. 

The distribution of these remarkable excretory tubules 
or nephridia is coextensive with that of the pharyngeal 
gill-clefts. They extend from the anterior to the posterior 
extremity of the pharynx, but not beyond this. They 
never have more than one opening into the atrial cavity, 
but those occurring in the mid-region of the pharynx have 
several, sometimes as many as nine, openings into the dor- 
sal coelom. The number of ccelomic openings decreases 
anteriorly and posteriorly, until, at the two extremities 
of the pharynx, there is only a single coelomic opening 
to the tubules. 

In a full-grown individual, Boveri has counted ninety- 
one tubules on one side of the pharynx, the total number 
therefore being double this. 

The serial distribution of the excretory tubules, one 
after the other, is known broadly as a metamcric arrange- 
ment. But since they correspond in number and situa- 


tion to the primary gill-clefts, which are much more 
numerous than the myotomes in the region of the body 
in which they occur, their arrangement is more strictly 
defined as branchiomeric. In the larva, however, the pri- 
mary gill-slits correspond numerically with the myotomes 
or muscle-segments of the pharyngeal region, only sec- 
ondarily becoming more numerous. The branchiomeric 
arrangement of the excretory tubules of Amphioxus need 
not, therefore, prejudice their claim to be regarded as 
segmcntal structures. 

If, now, we attempt to compare the nephridial system 
of Amphioxus with the kidney of the higher types, we 
shall find that here also, as in so many other instances, 
the permanent state of things in the former becomes a 
characteristic feature of the embryo in the latter. 

As is well known, the kidney of the higher Vertebrates 
comprises a mass of convoluted tubules, the uriniferous 
tubules, imbedded in a matrix of fibrous connective tissue, 
and enclosed within a common sheath, and so producing 
collectively a compact organ which we call the kidney. 

If, neglecting the highly elaborate structure presented 
by the kidney of Birds and Mammals, we take, as a typi- 
cal example of a primitive Vertebrate renal organ, that of 
a tailed Amphibian, we find after a superficial examina- 
tion the following characteristic features. In the newt, 
for instance, the surface of the elongated kidney is studded 
with numerous small apertures. These are surrounded by 
vibratile cilia, and lead directly from the body-cavity into 
the convoluted renal tubules. They are, therefore, the 
ccelomic openings or funnels of the latter, and are known 
as nephrostomes. Close to the nephrostome a short diver- 
ticulum of the tubule leads to a capsule which encloses a 
glomcrulus. After a winding course in the substance of 


the kidney, the tubules emerge from the latter as a series 
of efferent ducts placed one behind the other, and these 
again open into a common longitudinal duct on each side 
of the body, known as the ureter, which leads the products 
of excretion backwards to the cloaca. 

The permanently functional kidney of Fishes and Am- 
phibia is known as the mesonephros. In Reptiles, Birds, 
and Mammals, this is only functional during the embryonic 
period, and later is replaced in a way not yet fully eluci- 
dated by the permanent kidney of these forms which is 
known as the metanephros. 

The ureter, or duct, of the mesonephros, is spoken of as 
the mcsonepJiric duct, while the renal tubules constitute, 
collectively, the glandular portion of the kidney. 

The permanent kidney of the craniate Vertebrates is ab- 
solutely unique among all the other glands of the body, in 
the fact that the glandular portion of the organ arises 
independently of the duct, and only communicates secon- 
darily with it. Moreover, the duct develops in point of 
time before the gland. This is a very extraordinary fact, 
and taken alone would be quite inexplicable. It has been 
found, however, that the mesonephric duct has primary 
relations with a totally distinct set of excretory tubules, 
which differ from those mentioned above, both in their 
position in the body and in their mode of development. 
These primitive tubules, which mark the first appearance 
of a renal organ in the Vertebrate embryo, constitute the 

The degree of development attained by the pronephros, 
or primitive kidney, in the life-history of the various types 
of Vertebrates, is very different in the different classes. 

Frequently, as with the Selachians (sharks), Birds, most 
Reptiles, and with the Mammals, the pronephros is an 


entirely rudimentary structure, which puts in a fleeting 
appearance during the embryonic development, but never 
functions as a kidney. 

In other cases, as with the Teleostomes, or bony fishes, 
Amphibians, Crocodiles, and Turtles, the pronephric sys- 
tem attains a higher grade of development, and actually 
functions for a time as the sole kidney of the animal. In 
some of the bony fishes (e.g. Zoarces and Merlucins), it 
functions as the kidney for an extraordinarily long time, 
apparently throughout the period of adolescence. In one 
curious instance of a fish, Fierasfer, which has acquired a 
semi-parasitic habit, it appears that the development has 
been arrested to such an extent that the pronephros 
functions as the principal organ of excretion throughout 
life, the mesonephros remaining rudimentary (EMERY). 

The most extensive pronephric system which has as yet 
been described for any craniate Vertebrate, is that repre- 
sented diagrammatically in Fig. 32. This is the larval 
excretory system of a remarkable worm-like legless Am- 
phibian, Ichthyophis glutinosits, belonging to a very primi- 
tive subdivision of the Amphibia known as the Cceciliani, 
which occur in the hot regions of South America, Africa, 
Seychelles, East Indies, and Ceylon. 

We owe our knowledge of this elaborate pronephric 
system to RICHARD SEMON of Jena. 

It consists of some twelve pairs of irregularly contorted 
tubules placed dorsal to the general body-cavity in a posi- 
tion which is described as retro-peritoneal, and arranged seg- 
mentally, one behind the other, on either side of the dorsal 
aorta. Broadly speaking, the canals run outwards in a 
transverse direction. Near their inner extremities they 
usually divide into two short branches, which terminate 
each in a funnel-shaped opening into the body-cavity. 



Fig. 32. Pronephric system of embryo of Ichthyophis, reconstructed from sec- 
tions, and represented as having been spread out in one plane. (After SEMON.) 

a. Dorsal aorta, c. Portions of the ccelom into which the nephrostomes of the 
pronephric tubules open. The inner portion of ccelom (next to aorta) is shut off 
from the rest of the ccelom, and becomes associated with the vascular outgrowths 
from the dorsal aorta (which produce the glomeruli) to form the Malpighian cap- 
sules of the pronephros. The Malpighian tractus is continued backwards as 
a metamorphosed and rudimentary cord of cells, nearly to the cloaca, and con- 
stitutes the so-called Nebenniere or Interrenal body. This backward extension 
of the Malpighian body of the pronephros probably indicates the former existence 
of a much more extensive pronephric system, p. Convoluted pronephric tubules 
lying above the peritoneum (shaded light), each provided with two nephrostomes, 
inner and outer, and opening peripherally into d, the longitudinal pronephric duct 
(Wolffian duct), which becomes the mesonephric duct after the degeneration of 
the pronephric tubules and the formation of the mesonephric tubules have taken 
place, m. Rudiments of the mesonephric tubules. 

N.B. The pronephric tubules are here characterised by the possession of 
ccecal outgrowths. 


6 9 


These are the ccelomic openings, or nephrostomes, of the 
tubules. At their outer ends most of them open directly 
into a longitudinal duct, \kspronephric duct, which extends 
backwards to the cloaca. 
The most anterior tubules, 
however, tend to fuse to- 
o-ether at their outer ex- 


tremities, before reaching 
the common duct. Corre- 
sponding to each tubule 
there is a short artery 
growing out from the dor- 
sal aorta, and abutting with 
its blind end against the 
portion of the body-cavity 
into which the innermost 
nephrostomes open. 

Later on these coecal 
outgrowths from the dorsal 
aorta develop a vascular 
network at their free ends, 
and so produce a series of 

pig 23. Schematic transverse section 
If nOW we inquire into through a Selachian embryo in the region 

of the pronephros. (After VAN WlJHE.) 

the mode of development The dotted line drawn across the section 
,- . , indicates the plane of division between the 

of such a pronephnc sys- upper segme F nted and the lower unseg- 

tem as the One above de- rnented portions of the primitive body-cavity 

(procoelom). my. Myotome or myomere. 

Scribed, we find that its . Mesomere or nephrotome. p. Prone- 
, , phric outgrowth, sp. Unsegmented bodv- 

component tubules arise as avity or - splanchn ; coel , sdero.ome. 

a Series of knob-like Seg- Nerve-tube, ch. Notochord. ao. Dor- 
sal aorta, a/. Digestive tube. 

mental outgrowths from 

the outer or somatic layer of the peritoneum at the base 

of the segmented portion of the primitive body-cavity. 


These outgrowths are at first solid cell-proliferations of 
the peritoneal epithelium, in the midst of which a lumen 
is subsequently formed between the cells. As soon as 
this occurs, the peritoneal thickenings represent hollow 
diverticula of the ccelom, each communicating with the 
latter by a single nephrostome (Fig. 33). 

The incipient tubules then grow outwards until they 
reach the ectoderm with which, in the Selachians, they 
become fused. This has been taken by Ruckert to indi- 
cate that the tubules originally discharged the products 
of excretion directly to the exterior by a series of indepen- 
dent apertures at the points of fusion. (Cf. Fig. 34 A.) 5 
The pronephric tubules next commence gradually to relin- 
quish their coalescence with the ectoderm from before 
backwards, retaining, however, for the present the connec- 
tion behind (Fig. 34 B}. 

Meanwhile the distal ends of the successive tubules 
undergo confluence (Fig. 34 B), and in this way the begin- 
ning of a longitudinal duct is produced. This duct now 
gradually splits itself off from the ectoderm, so that the 
posterior connection with the latter is carried farther and 
farther back until it reaches the region of the cloaca, when 
it leaves the ectoderm and acquires an opening into the 
cloaca (Fig. 34 C). Meanwhile, however, in the Sela- 
chians, the pronephric tubules begin to undergo a retro- 
gressive development and atrophy, as a consequence of 
which the pronephros as a gland becomes aborted. 

In the same way, but at a much later stage, the remark- 
able pronephric system of Ichthyophis becomes entirely 
aborted. But the duct remains, and a new set of tubules 
appear at the bases of the somites, which secondarily open 
into it (Fig. 34 C). 

These new tubules are the mesonephric tubules, and, 


although they occur mostly behind the region of the pro- 
nephros, yet rudiments of them appear in the same seg- 
ments occupied by the latter. Unlike the pronephric 
tubules, they arise, not as evaginations from the base of 
the somites, but in such a way that an adjacent portion 
of the somite, lying dorsal to the pronephric tract, loses 



Fig. 34. Three diagrams illustrating the hypothetical phylogenetic develop- 
ment of the excretory organs in Selachians. (After RUCKERT.) 

s. Somites, pn. Pronephric tubules fused with ec, the ectoderm in A ; collected 
into a common duct w.d, the Wolffian or pronephric duct in B; and finally 
aborted in C, with the exception of one, which persists as the ostium abdominale. 
inn. Mesonephric tubules, w.d. Pronephric duct in B; mesonephric duct in C. 
cl. Cloaca, p. Posterior region. 

its primary connection with the rest of the somite, which 
consists of the myotome proper, and becomes bodily con- 
verted into a mesonephric tubule whose blind end curves 
round the pronephric duct and eventually opens into it ; 
while its point of communication with the unsegmented 


body-cavity persists as the nephrostome. (Cf. Figs. 33 
and 35 B.) 

The pronephric duct, therefore, becomes secondarily 
employed in the surface of the mesonephros. So that, 
while the mesonephros and its future duct form two dis- 
tinct morphological structures, the pronephros and the 
same duct form one inseparable whole. 

From the above considerations we may conclude that 
the pronephros represents the primitive and ancestral 
excretory organ of the craniate Vertebrates. Just as the 
notochord has been largely replaced first by cartilage and 
then by bone, so the pronephros has been replaced first 
by the mesonephros and then by the metanephros. 

Returning now to Amphioxus, we have to note in the 
first place the absence of a common matrix surrounding 
the excretory tubules, and, secondly, the absence of a com- 
mon duct. Since in the higher Vertebrates the interstitial 
growth of connective tissue among the tubules, binding 
them together into a compact organ, is a secondary phe- 
nomenon, the absence of such a matrix in Amphioxus 
need not detain us. 

Judging from the analogy of the other systems of or- 
gans in Amphioxus, it will be at once concluded that the 
excretory tubules of the latter represent the pronephric 
system of the embryos of the craniate Vertebrates. And 
this, in fact, is Boveri's contention. 

As we have seen, the excretory tubules of Amphioxus 
open separately into the atrial cavity. While they do not, 
therefore, open directly to the exterior at the ectodermic 
surface of the body, they do actually open at an ecto- 
dermic surface, since the atrial cavity is a space enclosed 
from the outside, and so is lined by ectoderm. The pri- 
mary fusion of the pronephric tubules with the ectoderm, 



which has been observed in some craniate Vertebrates as 
described above, is therefore probably of the same nature 
as the ectodermic openings of the tubules in Amphioxus. 

Fig. 35. A. Schematic transverse section through pharyngeal region of Am- 
phioxus. On the left is a branchial bar, cut lengthwise, and on the right a gill-slit. 

B. Schematic transverse section through Selachian embryo. (After BOVERI.) 

at.c. Atrial chamber, p.n.d. Pronephric duct. c.o. Nephrostome of pronephric 
tubule, k.t. Cross-section of excretory tubule in Amphioxus. a.f. Opening of 
excretory tubule into atrium in Amphioxus. g.c. Gonadic cavity (perigonadial 
coelom) in A ; compared by Boveri with the mesonephric tubule, mes.t. in B. 
gl. Glomerulus. ca>. Ccelom. e.c. Endostylar ccelom. s.i.v. Branchial artery in 
A ; sub-intestinal vein in B. 

Other letters as in previous figures. 

N.B. In B the future opening of the mesonephric tubule into the pronephric 
duct is indicated by dotted lines on the right. The vessel connecting the sub- 
intestinal vein with the aorta is placed on the left of the alimentary canal for com- 
parison with Fig. A. It is really only present on the right side, although a rudiment 
occurs on the left. (See Note 6.) 

The glomeruli of the tubules in Amphioxus are supplied 
by blood-vessels which connect the dorsal aorta with the 
branchial artery. It should be remembered that the bran- 
chial artery represents the anterior portion of the sub- 


intestinal vein, and in the young larva the two vessels are 
continuous. The direct continuity is subsequently inter- 
rupted by the development of the hepatic ccecura, and the 
consequent insertion of a capillary portal system into the 
circulation. In the Selachian embryo, a series of similar 
vessels, six in number, connecting the dorsal aorta with 
the sub-intestinal vein, have been shown to be in close cor- 
respondence with the pronephric tubules, and to form at 
the level of the tubules a series of rudimentary glomer- 
uli (Figs. 35 A and B)? 

Such resemblances as the above are demonstrative, and 
are sufficient to prove that the excretory tubules of Am- 
phioxus belong to the pronephric system, and that in this 
respect, also, the adult Amphioxus presents features which 
are characteristic of the embryos, or larvae, of the higher 

Although convinced as to the essential identity of the 
excretory tubules of Amphioxus with the pronephros of 
the craniate Vertebrates, it must be remembered that 
there is one apparently great difference between them. 
Whereas in Amphioxus the pronephros (applying this 
term to the tubules considered collectively) occurs in the 
region of the perforated pharynx, in all the higher Verte- 
brates it occurs behind the pharynx, and is quite absent 
from the regies of the gill-slits. This difference, however, 
which might at first sight appear serious, is, in reality, 
most instructive. As Boveri points out, it shows almost 
conclusively that the pharynx of Amphioxus does not 
correspond to the pharynx alone of the higher forms, but 
to the pharynx together with the anterior portion of the 
alimentary canal. 

In the Craniota the gill-clefts, which are present in a 
limited number, have becomfc involved in the complicated 


process of cephalisation, by which the Vertebrate head has 
been evolved. They are innervated exclusively by the 
cranial nerves, and in fact are considered as forming part 
of the head. In Amphioxus there is, broadly speaking, no 
head, and the region of the gill-slits forms part of the trunk. 
In the evolution of the Craniota, therefore, what has hap- 
pened is that the gill-clefts have been relegated to the 
head, while the excretory tubules have become confined to 
the trunk, and have ceased to occur in the neighbourhood 
of the gill-clefts. Only the anterior region of the pharynx 
of Amphioxus is represented by the pharynx of the higher 
forms. The greater part of it corresponds to the unper- 
forated portion of the alimentary canal, which follows 
immediately behind the pharynx in these forms, extending 
to the liver. 

We have referred above to the absence of a pronephric 
duct in Amphioxus. Although this is true in the strict 
sense of the term, yet Boveri gives reasons for supposing 
that the right and left pronephric ducts are in a measure 
represented by the right and left halves of the atrial 
chamber. (Cf. Fig. 35, A and B). We will first glance 
briefly at the mode of 

Development of the Atrial Cavity. 

For the sake of avoiding complications, it will be well to 
confine the description at present to the mode of origin of 
the atrial cavity in its posterior region. It arises of course 
on the same principle throughout its whole extent (except 
the post-atrioporal continuation, which ^ grows back later^ 
but anteriorly it is involved in the asymmetry which is such 
a marked feature of the larva, and will be considered in the 
chapter on the general development. 

The first indication of the fjiture atrial cavity appears in 

7 6 


a young larva with some six or seven gill-slits in the form 
of two longitudinal thickenings of the integument on the 
ventral surface of the body. These are at first solid, but 
eventually become hollowed out so as to enclose a longitu- 
dinal canal on each side. This is the so-called metapleural 
canal or lymph-space. The thickenings enlarge to the 
extent of forming two well-marked folds of the body-wall ; 
namely, the metapleural folds. 

The next stage is marked by the formation of two small 
solid longitudinal ridges on the inner opposed faces of the 
metapleural folds (Fig. 36). It is by the subsequent 

Figs. 36 and 37. Schematic transverse sections through post-pharyngeal 
region, illustrating mode of origin of atrial chamber. (After LANKESTER and 


ao. Aorta, b.c. Coalom. r.m and l.m. Right and left metapieural folds, s.a.r. Sub- 
atrial ridges, which fuse together to form the floor of at, the atrium, int. Aliment- 
ary canal, s.i.v. Sub-intestinal vein. 

meeting and coalescence of these subatrial ridges that the 
atrial cavity becomes enclosed as a small median tube lined 
by ectoderm. 

As soon as it has become closed off from the exterior, 
the atrial tube commences to grow in size, and it gradually 



expands laterally and also in an upward direction, propor- 
tionately reducing the extent of the ccelom as it does so 
(Fig. 37 ; cf. also Fig. 26). At its posterior extremity the 
atrial tube does not become closed in, but remains perma- 
nently open as the atriopore. 
It is a curious fact that the 
fusion of the subatrial ridges 
to enclose the atrial tube takes 
place gradually from behind 
forwards, so that for a long 
time the latter has the form 
of a canal open to the exterior 
at both ends. The chief feat- 
ures in the formation of the 
atrium are shown diagrammat- 
ically in Fig. 38, A, B, and C. 
In Fig. 38 A the atrial tube 
has not begun to be closed in, 
but the two metapleural folds 
are seen running side by side 
for some distance. Anteriorly 
the development of the right 

, . . , r , Fig. 18. Three plastic diagrams 

metapleur is in advance of that of lar ^ of Amphioxu ; from th * ven . 

of the left, and it is Seen tO tral aspect, illustrating the mode of 

enclosure of the atrial tube from be- 
bend round tO the right Side hind forwards. The atrium is still 

of the body in correspondence S2VS?S ail* U^S 

with the asymmetry of the gill- closed in C. (After LANKESTER and 


slitS (vide infra). Having ar- p. s . Primary gill-slits. r.m. Right 

rived at the front end of the Praoialpit " Moutht 

pharynx, the right metapleur 

bends sharply inwards to the mid-ventral line and then 
gradually dies out in front. In Fig. 38 B the subatrial 
ridges have met and fused for a short distance behind the 


pharynx, so as to enclose a tube which corresponds to that 
portion of the future atrial cavity which lies between the 
atriopore and the hinder end of the pharynx. Finally, 
in Fig. 38 C, the closure of the atrial tube has advanced 
forwards over the gill-slits almost to the anterior extremity 
of the pharynx, still leaving, however, one or two gill-slits 
open directly to the exterior in front. Meanwhile, the 
floor of the atrium has increased in width, and the meta- 
pleural folds are separated by a wider interval than before 
(Fig. 38 C}. Eventually the atrium closes up completely 
in front, so that the gill-slits no longer open directly to 
the exterior. 

Remembering that the atrium of Amphioxus arises as an 
unpaired median tube (see below, IV.), while the pro- 
nephric duct is always paired, the following are some of 
the reasons for supposing a partial homology between the 
two structures :- 

(a) They are both derived, either wholly (atrium), or in 
a large measure (pronephric duct), from the ectoderm. 5 
(ft) They both receive and carry away the excretory prod- 
ucts from the pronephric tubules ; and (7), they are 
both, to a greater or less extent, lined by an epithelium, 
which is itself glandular and excretory." 

Comparison between the Excretory System of AmpJiioxns 
and that of the Annelids. 

Having considered the relation existing between the 
pronephric system of Amphioxus and the corresponding 
system in the embryonic and larval stages of the higher 
Vertebrates, we will now pass on to a brief comparison 
with the excretory system of the Invertebrates. 

The excretory system of a typical Annelid presents 



certain resemblances to that of Amphioxus, in that it 
occurs in the form of distinct segmental tubules, or 
nephridia, each possessing a funnel-shaped opening into 
the body-cavity, and an opening to the exterior at the sur- 
face of the body. 

It was, in fact, the recognition, some twenty years ago, 
by SEMPER and BALFOUR, of the resemblance between the 
arrangement of the nephri- 
clia of the Annelids and 
the primary segmental ori- 
gin of the kidney of the 
Craniota that was chiefly 
instrumental in placing the 
Annelid-theory of Verte- 
brate descent on a tempo- s ; 
rarily firm basis. 

A dissection of the an- 
terior portion of the body 
of an earthworm, exposing pig 39 _ Anterior portion of earth . 

the nephridial tubules, is worm dissected open from above to show 

the nephridia and nervous system. (From 

shown in Fig. 39. A pair \y. T. SEDGWICK and E. B. WILSON'S 

of such convoluted tubules m (prseoral lobe) . ^ 

OCCUrS in each segment, Or Cerebral ganglion, which has receded from 

the prostomium from the ectoderm of 
ring, Of the body, COm- w hich it arose, com. Circumoesophageal 

menHno- from the third commissure surrounding the buccal tube 

"S (latter not represented). v.n.c. Ventral 

Physiologically, of COUrse, nerve-cord. . Segmental nerves. nph. 

Nephridia. sp. Dissepiments. 

they are directly com- 

parable to the renal tubules of the Chordata, and in their 
general features, allowing for the absence of a common 
duct, the similarity in the two cases is striking enough. 
But when this undoubted similarity is used as an argument 
for deriving the Vertebrate excretory system directly from 
that of the Annelids, we tread on very uncertain ground. 


If we were to consider the excretory system apart from 
the rest of the organisation, this would be the only course 
to follow. But when the whole organisation is taken into 
account, the only justifiable conclusion seems to be, not 
that the Vertebrate renal system is to be derived from that 
of the Annelids, but that, as Ruckert suggests, both may 
possibly have been evolved from a common starting-point. 

It is eminently probable that, in respect to this and the 
other systems of organs, as well as the segmentation of 
the body, the Annelids and Vertebrates present an in- 
stance of parallel evolution. This will become more evi- 
dent as we proceed. Those who uphold the so-called 
Annelid-theory have no cause to complain of the absence 
of a common duct to the nephridia, since this has been 
found in some cases to occur. 

In 1884 EDUARD MEYER discovered that in certain 
marine Annelids (Lanicc concJiilega and Loimia medusa) 
belonging to the family of the Terebellidas, the nephridia 
of each side were joined together by longitudinal ducts, 
which he compared, though with great reserve, to the 
mesonephric ducts of the Vertebrata.* In these worms the 
nephridia do not occur in all the segments of the body, but 
are confined to the anterior so-called thoracic region, their 
number being very limited. In the thorax, the dissepi- 
ments which typically divide the segments from one 
another are absent, so that the body-cavity would here 
form a continuous uninterrupted space, were it not that it 
is divided into two chambers, an anterior and a posterior, 
of which the latter is the larger, by a muscular diaphragm. 
In the anterior thoracic chamber (Fig. 40) there are three 
pairs of nephridia which are united together on each side 
by a short duct opening to the exterior by a single aperture. 

* This discovery was also made later but independently by J. T. CUNNING- 
HAM for Lanice conchilega. 




In the posterior chamber there are four pairs of much 
larger nephridia, which are similarly joined together by a 
prominent longitudinal duct from which short processes 
corresponding in number to the nephridia lead to the 
external apertures. The 
duct itself ends blindly at 
both ends, but is prolonged 
posteriorly far beyond the 
region of the nephridia 
(Fig. 40). 

The presence of this 
longitudinal duct in these 
worms is a very remark- 
able circumstance, but it is 
undoubtedly an expression 
of the same phenomenon as 
the anastomoses between 
successive nephridia which 
have been described by 
EISIG for the Capitellidae, 
as well as the complicated 

Series of anastomoses which Fig. 40. Schematic lateral view of 

. anterior end of Lattice conchilega to show 
Convert the entire nephn- the nephridia. (After EDUARD MEYFR 

dial system into a marvel- fro Hatsche k's LehMj derZooio^ 

J The ventral side of the body is to the 

loilS network Of tubules dis- left of th e figure, d. Longitudinal ducts of 

11 A C* T3 *^ e ne Ph r 'dia. e.o. Position of external 

Covered by A. (j. -BOURNE openings. / Nephridial funnel ( = coalomic 

in the marine leech Pon- P enin s of nephridium). /. Position of 

mouth ; bounded by two prominent lateral 
tobdella, and by BEDDARD lobes, and fringed by a great number of 

" feelers," which are cut short in the figure, 
in the CUriOUS earthworm, ,. Branchial tentacles (three on each side 

Perichceta. of the body). 

The present state of our knowledge does not admit of 
an attempt to specify the particular type of nephridial 
system from which that of the Annelids, on the one hand, 


and that of the Vertebrates, on the other, took their 


In view of the apparent absence of nephridial tubules 
in Balanoglossus and the fact that in the Ascidians the 
renal organs are special structures peculiar to this group, 
it is extremely difficult to associate the Vertebrate type 
of excretory system with that of any Invertebrate. 

Since the Annelid-theory precludes the possibility of 
Amphioxus being regarded as an ancestral form, and yet 
if, nevertheless, it is, as we believe, primitive and not 
essentially degenerate, the discovery of the excretory 
tubules in Amphioxus happily releases us not only from 
necessity, but also from the possibility of referring the 
Vertebrate excretory system back to that of the Annelids. 

Nervous System. 

The central nervous system of Amphioxus consists of a 
closed thick-walled tube lying along the dorsal side of the 
body above the notochord. 

Viewed externally, it is a perfectly plain, more or less 
cylinder-shaped structure, without any constrictions or 
enlargements whatever. Its largest diameter in the adult 
occurs about the middle of its course, and not at its 
anterior end. 

Posteriorly it is nearly coextensive with the notochord, 
and, like it, tapers down almost to a point.* Anteriorly it 
terminates abruptly some distance behind the front end 
of the notochord. (Cf. Figs. 3 and 11.) 

If the dorsal nerve-cord be removed from the body and 

* The extreme posterior end of the nerve-cord is usually swollen out into 
a small ampulla-like dilatation. (PouCHET, RoHON, RETZIUS.) RETZIUS 
has observed that occasionally the nerve-cord is prolonged beyond the dilata- 
tion and actually bends round the posterior end of the notochord. 


examined from above, its general appearance will be as 
shown in Fig. 41. In front there is a pair of nerves 
which proceed symmetrically from the sides of the nerve- 
tube. Farther back there is 
another pair of nerves which 
arise more dorsally than the 
anterior pair, but are likewise 
placed symmetrically one oppo- 
site the other. Behind this 
second pair of nerves the spinal 
nerve-roots are no longer dis- 
posed symmetrically, but alter- 
nate with one another, in cor- 
respondence with a sirriilar 
alternation of the myotomes, 
the alternation becoming more 
and more pronounced as we 
proceed backwards. Again, be- 
hind the second pair of nerves 
there are two kinds of spinal 
nerve-roots, dorsal and ventral. 
The former leave the nerve-cord 
from its dorsal surface, and the 
latter from the margins of its spi^'co 4 ^' ~ 
ventral side. In the dorsal roots from above. (After SCHNEIDER.) 

Between the first pair of cranial 
the nerve-fibrils are Collected nerves is seen the eye-spot ; one of 

together to form a single com- the br ^h es of the second pair of 

cranial nerves sometimes arises 

pact nerve round which the directly from the spinal cord as 

r shown on the right; farther back 

Sheath Ot the nerve-COrd IS COn- are seen the pigment spots of the 

tinned, while in the ventral roots nerve - cord - 
the nerve-fibres emerge separately in loose bundles unsur- 
rounded by a sheath, from the spinal cord. A pair of 
dorsal roots and a pair of ventral root-bundles go to each 


segment of the body. Dorsal and ventral roots are entirely 
independent of one another, and at no point do they coa- 
lesce as they do in the Craniota. In further contrast to 





Fig. 42 A.-- Innervation of the region of the oral hood and snout. (After 
HATSCHEK, slightly altered according to the statements of VAN WIJHE.) 

ck. Anterior end of notochord. ci. Buccal cirri. cn l , en' 1 . First and second 
cranial nerves with their peripheral ganglia, at. Rami cutanei dorsales. l.k. Left 
half of oral hood. r.h. Right half of oral hood. o. Olfactory pit. s/>\ sf>*. First 
and second dorsal spinal nerves, so. Sense-organ of oral hood (groove of Hat- 
schek) indicated as if seen through body-well by transparency, v. Velum. 
v.n.1. Nerve to left side of velum. v.n.r. Nerve to right side of velum. 

N.B. The septa between the myotomes are indicated by dotted lines. The 
superficial nerves of oral hood are rendered in black; the deeper nerves, which 
anastomose to form the plexus of Fusari, are left white. 

the condition met with in the latter there is no ganglionic 
enlargement on the dorsal root. 


The first two pairs of nerves differ in many points from 
those which succeed them, and are known as the cranial 
ncrres. Thus they have no corresponding ventral roots ; 
they appear to be exclusively sensory, and do not inner- 
vate any muscles ; their distribution is confined to the 
snout, and they are above all characterised by the pres- 
ence of peripheral ganglionic enlargements which occur 
chiefly on the finer branches of 
the nerves near their distal ex- 
tremities. Furthermore they lie 
in front of the first myotome. 
The first pair of dorsal spinal 
nerves (i.e. the third pair alto- 
gether) belonging to the first 
myotome passes from the nerve- 
tube to the skin through the 
dissepiment which separates the 

Fig. 42 B. Diagram illustrat- 

first myotome from the Second, ing the branching of a dorsal spinal 
A i ,, n ,1 i- nerve of Amphioxus. (After HAT- 

And so with all the succeeding SCHEK) 

dorsal roots, they lie at the back d -''- Dorsal root - *<* Ramus 

dorsalis. r.v. Ramus ventralis. 
Of the myotome tO which they r .vi. Ramus visceralis. r.c. Ramus 

1 i i , i ,1 cutaneus ventralis innervating ecto- 

belong, between it and the next derm of metapleun ,. | entral 

followin " Segment. (Cf. Fi"S. or motor root, indicated as if in the 

same plane as the dorsal root. 

2 and 42 A.) 

Shortly after leaving the central nervous system, the 
dorsal roots divide into two branches, a minus dorsalis 
and a ram us ventralis (Fig. 42). These two branches run 
upwards and downwards respectively, in the gelatinous 
layer of the sub-epidermic cutis ; that is to say, external 
to the muscles. 

In the Craniota the corresponding branches of the 
spinal nerves lie for the first part of their course internal 
to the muscles, between the latter and the notochord. The 


cranial nerves of the Craniota so far resemble the dorsal 
spinal nerves of Amphioxus that they run external to or 
ectad of the somites of the head. 

The ramus dorsalis of a spinal nerve breaks up into a 
number of finer nerves, which supply the skin of the back. 
The ramus ventralis similarly gives rise to a number of 
cutaneous nerves, but in addition it gives off a branch 
which passes inwards below the longitudinal muscles of 
the body-wall, between them and the transverse muscles 
which lie in the floor of the atrium. This is the visceral 
branch of the spinal nerve. The visceral nerves innervate 
the transverse muscles and form an elaborate plexus on 
the surface of them.* 

Thus the dorsal spinal nerves of Amphioxus are of a 
mixed nature, sensory and motor, but chiefly sensory. 

The ventral roots are entirely motor. On their emer- 
gence from the spinal cord they spread out like a fan 
and terminate upon the muscle-fibres of the myotomes 
(Fig. 43). 8 

The muscles which are not innervated by the ventral 
spinal nerves are the transverse or subatrial muscles, the 
muscles of the montli (velum), and oral hood, and probably 
the anal sphincter. These are supplied by the so-called 
visceral branches of the dorsal nerves. The nerve-supply 
of the oral hood is illustrated in Fig. 42. It arises from 
branches of the third to the seventh dorsal nerves. These 
branches are distributed in two different ways : one set 

* The visceral nerves also send up branches, which pass up through the 
ligamentum denticulatum to the wall of the pharynx. (FuSARl; see below, 
p. .) Here they form the branchial plexus described by RoHON, who 
thought these nerves contained elements of the J'agus of the Craniota. 
The portions of the visceral nerves innervating the transverse muscles (these 
branches being discovered by ROLPH) were held by ROHON to contain 
elements of the Sympathetic system of Craniota. 


of them runs beneath the outer surface of the oral hood 
and, by the occurrence of frequent anastomoses, forms a 
coarse network known as the outer plexus, while the other 
set lies beneath the inner surface of the oral hood and 
gives rise to the inner plexus. The latter was discovered 
by FUSARI in 1889. The two plexuses are distinct from 

Fig. 43. Transverse section through the spinal cord in the middle region of 
the body. (After ROHDE. ) 

a. Giant fibre proceeding from the giant ganglion-cell A (see below), c.c. Cen- 
tral canal, g.f 1 . Giant nerve-fibres, which traverse the spinal cord from before 
backwards, g-f' 1 . Giant fibres, which traverse the spinal cord from behind for- 
wards, m.p. Muscle-plates, m.r. Motor nerve-fibres, n.f. Longitudinal nerve- 
fibres cut across, s.f. Supporting fibres, sh. Sheath of nerve-cord ( = dura mater ; 

one another, except in so far as their component nerves 
have a common origin from the dorsal roots (Fig. 42). 
The outer plexus is continued up into the individual cirri, 
while the inner plexus appears to stop short at the base 
of the cirri. It has recently been discovered by VAN 



WIJHE that the inner plexus on both right and left halves 
of the oral hood is exclusively formed by nerves which 
arise from the left side of the central nervous system ; 
and, further, that the nerve-supply of the velum is fur- 
nished by branches from the fourth, fifth, and sixth dorsal 
nerves of the left side only. This asymmetrical innerva- 
tion of the velum and inner (glandular) surface of the 

oral hood will be referred to 
again after the consideration 
of the larval development. 

The peripheral ganglionic 
enlargements which are so 
characteristic of the two pairs 
of cranial nerves must be cor- 
related with the sensibility of 
the snout. As the nerve-fibres 
are continued beyond them, 
they are not to be regarded as 
end-organs, but simply as peri- 
Fig. 44. - Peripheral ganglion- pheral ganglia. Their structure 

cells of the cranial nerves of Amphi- j g snO wn in Fi^ 44 Thev 
oxus. (After FUSARI.) 

were discovered by the great 

French naturalist OUATREFAGES in 1845. Each of them 
is composed of from one to four nerve-cells, with granular 
protoplasm and a large nucleus. Each group is enclosed 
in a sheath which is a continuation of the sheath of the 
nerve itself. The sheath is lined internally by an endo- 
thelium. According to FUSARI the nerve-fibres enter into 
direct connexion with the cells, though some would appear 
to pass round them. 

The peripheral nervous system of Amphioxus can only 
be compared definitely, at present, in its broadest features 
with that of the higher Vertebrates. The determination 


of the particular homologies in the two cases forms one of 
the most difficult problems of comparative morphology. In 
correlation with the low grade of cephalisation to which 
Amphioxus has attained, there are only two pairs of 
cranial nerves, the succeeding nerves retaining their 
primitive spinal character. The dorsal spinal nerves, 
furthermore, possess features which are specially charac- 
teristic of the cranial nerves of the Craniota. Such are 
their mixed sensory and motor functions, and the position 
of their dorsal and ventral branches ectad of the muscula- 
ture. As already indicated above, the walls of the gill-slits 
of the craniate Vertebrates are innervated by cranial 
nerves, while in Amphioxus this is done by spinal nerves. 
(Cf. Fig. 92 ; see also below, p. 163.) 

In transverse section the spinal cord of Amphioxus is 
seen to have somewhat of a triangular shape. The central 
canal has the form of a vertically elongated split, commenc- 
ing from the vertex of the triangle, and extending two- 
thirds of the way downwards into the cord. For the 
greater part of its extent, however, the two sides of the 
canal are closely approximated together so as to obliterate 
the lumen, which widens out again below, and presents the 
appearance of a circular or oval tube. The sides of the 
canal are lined by an epithelium the cells of which, starting 
from an indifferent condition in the embryo, have become 
modified in several different directions. Some are ganglion- 
cells, and others send out long radial processes which trav- 
erse the substance of the nerve-cord, and serve to hold it 
together. These are the supporting fibres (Fig. 43). The 
cells in the nerve-cord form a much smaller proportion of 
the bulk of it than the nerve-fibres do. The latter run 
mostly in a longitudinal direction, and produce a punctate 
appearance in cross-section. 


Anteriorly in the region of the cranial nerves the lumen 
of the central canal widens out into a relatively spacious 
vesicle, known as the cerebral vesicle (Fig-. 45). In young 
individuals this cavity opens by an aperture called the 
neuropore into the base of an epidermal pit, which we 
have already described under the name of the olfactory 
pit. Later on the neuropore closes up, but its former 

- 45- -4- Brain and cranial nerves of a young Amphioxus of 3 mm. length. 
B, C, D. Sections through different portions of brain : B, through neuropore and 
cerebral vesicle ; C, through the intermediate portion, and D, through the dorsal 
dilatation of central canal. (After HATSCHEK.) 

ch. Notochord. c.v. Cerebral vesicle, dil. Dorsal dilatation (Hatschek's Fossa 
rhoinboidalis). e. Eye-spot, up. Neuropore. olf. Olfactory pit. 

/, //. First and second cranial nerves. 

presence is indicated by a shallow groove at the base of 
the otherwise solid stalk connecting the olfactory pit with 
the roof of the brain. 

Behind the cerebral vesicle the lumen of the central 
canal widens out in its dorsal portion independently of 


the ventral tube, so as to form a vesicular dilatation cov- 
ered over by a thin membrane. The region of the nerve- 
tube, over which this dorsal dilatation extends, has been 
compared by HATSCHEK, who discovered it, to the medulla 
oblongata of the craniate Vertebrates, which is similarly 
roofed in only by membrane. In the fully grown condi- 
tion, however, it seems to be largely obliterated by the 

Fig. 46. Transverse section through the spinal cord between the second and 
third sensory roots. (After ROHDE.) 

g.c. Dorsal aggregation of ganglion-cells (extending between the second and 
fifth pairs of sensory nerves ; a somewhat similar group of ganglion-cells occurs on 
ventral side of nerve-cord below the central canal between the fourth and sixth 
sensory nerves.) 

d.r. Dorsal root. s.f. Supporting fibres, c.c. central canal ; in this case equally 
wide throughout its entire height, and so all along the spinal cord. s/i. Sheath of 

development of a mass of large ganglion-cells which ex^ 
tend backwards as far as the fifth pair of sensory nerves 
(Fig. 46). 

All there is of a brain in Amphioxus is shown in Fig. 
45. The cerebral vesicle is a plain cavity without any 
true subdivision into ventricles. 9 In the development of 

9 2 


the central nervous system of the higher Vertebrates, a 
stage is passed through which may be compared broadly 
with the permanent condition of things in Amphioxus. 
But in the former the anterior portion of the medullary 
tube quickly becomes greatly enlarged in contrast to the 
spinal cord proper, and becomes divided by constrictions 
into fore-, mid-, and hind-brain, which constitute the three 
primary divisions of the Vertebrate brain. Then the 
brain undergoes a flexure round the anterior end of the 
notochord. This curvature of the primitively horizontal 
brain-region in the craniate Vertebrates is known as the 
cranial flexure. (Cf. Figs. 23 and 24.) 

Among the numerous longitudinal nerve-fibres which 
compose the bulk of the spinal cord of Amphioxus, there 

are some which stand out in 
marked contrast to the great 
majority on account of their 
large size. These are the so- 
called giant-fibres, and they form 
one of the greatest peculiarities 
in the spinal cord of Amphioxus. 
According to ROHDE there are 
no fewer than twentv-six of these 


giant-fibres present, and each of 

Fig. 47. Transverse section 

through spinal cord in region them arises from a correspond- 

of giant ganglion-cell G. (After f a-anglion-Ccll, These 

R.OHDE. ) o o o o 

a. Process of giant-ceil A. g.f. so-called giant-cells have many 


processes, i.e. they are multi- 
polar, but they each send out one main stem, which is a 
cnant-fibre. The giant-cells lie across the middle of the 

O O 

central canal, and the giant-fibres pass outwards alter- 
nately to the right or left of the central canal, and then 
bend downwards and pass below the central canal and up 



to the opposite side of the canal, where 
they continue their course in the longitu- 
dinal direction (Fig. 47). The giant-fibre 
belonging to the most anterior giant-cell 
differs in several respects from the other 
giant-fibres. It is much larger than the 
others, and, whereas the latter lie on either 
side of the nerve-cord, the fibre in question 
lies in the middle line immediately below 
the central canal (Figs. 43 and 47). 

These giant-fibres traverse the spinal 
cord almost throughout its entire length, 
stopping short at some distance from its 
anterior and posterior ends. The giant- 
cells are arranged one after the other in 
two groups, one group lying in the anterior 
third of the spinal cord, the fibres from 
which run backwards, and the other group 
occupying the posterior third of the cord, 
the fibres from which run forwards (Fig. 

The giant-fibres are in no direct con- 
nexion with the outgoing nerves, but the 
giant-cells usually occur opposite a sensory 
(i.e. dorsal) root (Fig. 49). 

In the spinal cord of Petromyzon giant- 
fibres are present in considerable numbers, 

Fig. 48. Scheme illustrating the course of the giant- 
fibres and their origin from the giant-cells A-7. in the spinal 
cord of Amphioxus. (After ROHDE.) Fl &- 4 8 - 

A-L. Giant ganglion-cells whose giant processes traverse the spinal cord from 
before backwards. A is about at the level of the sixth sensory root, counting from 
the first cranial nerve. M-7.. Giant ganglion-cells whose giant processes traverse 
the spinal cord from behind forwards. M is about at the level of the fortieth sen- 
sorv root. 



while in the higher Fishes and tailed Amphibia, as well 
as in the tadpoles of the anourous Amphibia, the giant- 
fibres are represented by the so-called fibres of Mauthner* 

They are not found in the spinal cord of adult tailless 
Amphibia, Birds, and Mammals. 10 

Their occurrence in such large numbers in Amphioxus 
is therefore the symbol of an archaic organisation. 

Giant-fibres form a very striking feature in the ventral 
nerve-cord of many Invertebrates. Here, however, they 

Fig. 49. Part of spinal cord seen from above; from a preparation stained 
with methylene-blue. (After RETZIUS.) 

g.c. Giant ganglion-cell lying across central canal, mo. Motor root. s. Sensory 

appear often to lose their nervous function, and serve 
rather as elastic supporting rods for the nerve-cord. They 
are enclosed in thick sheaths of connective tissue, and 
have been found to originate in giant ganglion-cells. 
When the enclosed nerve degenerates, they become hol- 
low tubes containing a coagulable fluid. (EisiG.) 

With regard to the internal origin of the nerves which 
pass out from the spinal cord, our knowledge only extends 
to the dorsal roots. At the base of the ventral roots the 

* Also known as Miillerian fibres. 



fibres appear to stop, and in their place a peculiar granular 
structure of unknown significance is found (Fig. 49). 

The fibres which constitute a dorsal root are derived 
from two sources. Part of them are continuations or 
branches of the longitudinal fibres on the same side of the 
nerve-cord, on which a given dorsal root may be, while the 
other moiety appears to arise largely from groups of small 
bipolar ganglion-cells in the neighbourhood of the central 

Fig. 50. Diagram illustrating the internal origin of the nerve-fibres of a sen- 
sory root. (Combination of two figures of RETZIUS.) 

The cells giving rise to the processes lying on the same side as a sensory root 
.S, which divide into a T at the base of the root, are naturally in contiguity with 
the central canal, but are displaced for the purpose of the diagram, m.l. Middle 

canal, which send one process each in the direction of the 
dorsal root, and another process from the opposite pole of 
the cell to join in with the longitudinal fibres of the other 
side of the spinal cord (Fig. 50). n 

We will now compare, or rather contrast, the central 
nervous system of Amphioxus with that of an Annelid 
such as the common earthworm. The type of nervous 
system presented by the latter is common to a vast propor- 
tion of the Invertebrates. It consists essentially of three 


very sharply defined parts (Fig. 39) ; namely, (i.) the cerebral 
or supracesophageal ganglion, which is situated dorsally 
over the buccal cavity; (ii.) a longitudinal solid nerve-cord 
composed of two more or less distinct halves, running 
along the whole length of the ventral side of the body 
below the alimentary canal ; (iii.) the circumcesopJiageal 
nerve-ring or commissure which encircles the buccal tube 
and connects the cerebral ganglion with the snbo2sophageal 
ganglion at the anterior extremity of the ventral nerve- 

Viewed from above (as in Fig. 39), the ventral nerve- 
cord presents a series of constrictions which are in some 
forms very pronounced. The wider portions occur in the 
middle of the body-segments, and constitute the ventral 
ganglia, which are strung together by the intervening 
nerves (connectives) in the form of a ganglionic chain. 
From the ganglia, paired nerves pass out to the organs of 
the body. 

One of the greatest peculiarities in the type of nervous 
system above described lies in the fact that the alimentary 
canal passes through and is surrounded by a portion of 
the central nervous system ; namely, the circumoesophageal 
commissure. This fact has been one of the most serious 
difficulties which the upholders of the Annelid-theory have 
had to contend with. 

In the Chordata the alimentary canal does not pierce 
the central nervous system in any sense whatever.* Never- 
theless, there have been many conjectures as to a possible 
equivalent of the circumoesophageal nerve-collar in the 
Vertebrates, although it is safe to say that nothing of the 
kind really exists. 

* Balanoglossus might be said to offer an exception to this rule (see 
Chap. V.). 



The ventral nerve-cord of the Annelids is no doubt in 
part physiologically equivalent to the spinal cord of the 
Vertebrates ; but since the two structures lie on opposed 
sides of the body, it is difficult to regard them as morpho- 
logically equivalent. Those who defend the Annelid-theory 
have postulated the occurrence of a half-revolution of the 
body in the supposed Annelid-like ancestors of the Verte- 
brates, as a result of which they acquired the habit of per- 
forming their locomotion, perhaps swimming, on their backs 
so that the ventral surface was turned uppermost. In this 
way, we are to suppose the original dorsal and ventral 
surfaces became reversed. This phylogenetic acrobatic 
feat with all its consequences is hard to imagine, and 
there are other alternatives which make it an unnecessary 
assumption. (See below, V.) 

The chief fundamental differences between the dorsal 
spinal cord of Amphioxus and of Vertebrates generally, 
and the ventral ganglionic chain of the Annelids, may be 
summed up as follows :- 






Nerve-cord is hollow. 
" dorsal. 

" " unconstricted. 
" " single. 

Ganglion-cells lie inside the 
fibrous layer. 


Nerve-cord is solid. 

" " ventral. 

" " constricted. 

" " double. 

Ganglion-cells lie outside the 
fibrous layer. 

As for the resemblances, in both cases nerves are given 
off segmentally, and also giant-fibres are present, whose 

function, however, is apparently very different in the two 

cases. J 

A R 



1. (p. 49.) LANKESTER has made the suggestion that there 
are not distinct capillaries and ccelomic space around the hepatic 
coecum, but that the space itself is capillariform. This view is in 
accordance with what one observes in transverse sections. 

2. (p. 50.) The fullest account of the contractile blood- 
vessels of Amphioxus, as observed in the living animal, is that 
given by JOHANNES MULLER. He observed the peristaltic con- 
tractions of the branchial artery (which is filled with a perfectly 
colourless blood), beginning from its hinder end, where it is joined 
by the hepatic vein (which also undergoes peristaltic contraction 
from before backwards along dorsal side of coecum) and extend- 
ing to the front end of the pharynx. The intervals between the 
successive contractions last about a minute. Immediately suc- 
ceeding upon the contraction of the branchial artery, the bulbils, 
which occur at the base of the primary or^forked gill-bars, contract 
too. He says that the heart-like "aortic arch" which occurs to 
the right of the velum (he thought there was one on the left side 
as well) contracts from below upwards, and that its contraction 
enabled him to discover it. As mentioned in the text, van Wijhe 
states that it has no communication with the branchial artery. 
Johannes Mliller also observed the peristaltic contraction of the 
sub-intestinal (portal vein), and states that it extends to the 
anterior end of the coecum. It should be remembered that his 
observations were made on young transparent individuals, and the 
statement as to the extent of the contraction of the sub-intestinal 
vein is open to doubt. 

3. (p. 51.) A genital artery running longitudinally above the 
gonadic pouches has been figured by Langerhans, Rolph, Schneider, 
Lankester, and Boveri, but its relations to the rest of the vascular 
system have not been made out. It is doubtful whether its 
presence is constant. 

4- (P- 58.) The "brown funnels" were discovered by LAN- 
KESTER in 1875, an d were subsequently compared by BATESON with 
the collar-pores of Balanoglossus. (See Chap. V.) This com- 
parison was made on the supposition that the posterior free oper- 



cular fold of the so-called collar in Balanoglossus is of the same 
nature as the atrium of Amphioxus ; but this is somewhat doubtful. 

5. (p. 70.) For an admirable critical and historical account 
of our knowledge of the development of the excretory system in 
the different groups of Vertebrates, the reader may be referred to 
the report on the " Entwickelung der Excretionsorgane" by Pro- 
fessor RUCKERT, in Merkel and Bonnet, Ergebnisse der Anatomic 
uiiii Entwickelirngsgeschichte, Band I., 1891. It will be sufficient 
to note here that the ectodermic origin of the pronephric duct, 
as briefly described in the text, only holds for the Selachians and 
Mammals. It was first discovered in the latter by GRAF SPEE in 
1884, and confirmed later by F LEMMING. In the former it was 
discovered independently by VAN WIJHE and RUCKERT (1886-8). 
On the contrary, in Petromyzon, Amphibia, Reptiles, and Birds, the 
duct does not arise from the ectoderm. 

Van Wijhe denied the segmental fusions with the ectoderm of 
the pronephric tubules in Selachians as described by Riickert. 
The account given by the latter author has, however, been 
indirectly confirmed by the observations of FELIX on the chick, 
where the pronephric outgrowths were found in some cases to 
undergo a transitory fusion with the ectoderm. 

BOVERI has attempted to show how the origin of the pronephric 
duct can be imagined to have been gradually transferred from the 
ectoderm to the mesoderm. Finally, it may be noted that, 
whereas RUCKERT compared the pronephric tubules with the 
Annelid nephridia, SEMPER and others employed the mesonephric 
tubules for the comparison. The fallacy of the latter comparison 
was first pointed out by FURBRINGER. 

6. (p. 74.) In 1887 PAUL MAYER discovered that the sub- 
intestinal vein in the Selachian (Pristiurus) embryo communicated 
with the dorsal aorta, by a series of six segmental vessels which 
passed up around the intestine on the right side only. Correspond- 
ing to them on the left side he found short, blind outgrowths from 
the dorsal aorta similar to those figured in the text in connexion 
with the pronephros of Ichthyophis. Paul Mayer's connecting 
vessels soon become aborted with the exception of one which 
enlarges and forms the proximal portion of the umbilical artery. 
In the following year it was shown in a brilliant manner by 


RUCKERT that these vessels occur in the same segments as the 
rudimentary pronephric tubules, and give rise to rudimentary 
glomeruli at the level of the tubules. (Cf. Fig. 35 .) There 
can be no doubt that these vessels are homologous with the 
vessels which run through the primary branchial bars of Amphi- 
oxus, and, as shown by BOVERI, assist in forming glomeruli at the 
level of the excretory tubules. 

The morphological importance of these facts is very great and 
has been strongly emphasised by Boveri. Whether Paul Mayer's 
connecting vessels indicate the former existence of gill-slits in that 
region is not so certain, since it is difficult to decide whether 
the indefinite number of gill-slits in the adult Amphioxus is a 
palingenetic (ancestral) feature or not. It should also be remem- 
bered that Paul Mayer found numbers of connecting vessels, 
between sub-intestinal vein and dorsal aorta, in the tail. 

7. (p. 78.) Boveri found that the epithelium of the pronephric 
duct of Myxine was of a glandular nature, comparable in this 
respect to the atrial epithelium of Amphioxus. 

8. (p. 86.) As shown in Fig. 43, ROHDE was inclined to 
follow SCHNEIDER in the belief that the fibres of the ventral spinal 
nerves were directly continuous with the muscle-plates and, more- 
over, exhibited the same striation as the latter. It has recently 
been shown by GUSTAV RETZIUS that this appearance of continuity 
is an illusion, as in so many other cases where nerves have been 
wrongly supposed to enter into direct continuity with peripheral 
end-organs. By employing Ehrlich's method of staining nervous 
tissue, infra vitam, with methylene blue, Retzius has proved that 
the motor fibres of Amphioxus pass with a somewhat winding course 
between the muscle-plates, and simply end on the surface of the 
plates. Rarely they branch dichotomously, but there is no special 
end-apparatus as in the higher forms. Their connexion with the 
muscle-plates is, therefore, one of intimate contiguity, but not of 

9. (p. 91.) The cerebral vesicle of Amphioxus was discovered 
thought it represented the fourth ventricle of the vertebrate brain. 
STIEDA (1873) was the first to homologise the cerebral vesicle of 
Amphioxus with the entire brain of the higher forms, and to regard 



it as representing the latter in its simplest form without any trace 
of subdivision. This view has very generally been adopted. Stieda 
also recognised the dorsal and ventral groups of ganglion-cells (of 
which the former is shown in section in Fig. 46) as belonging to 
the hinder portion of the brain. Rohde's conception of the brain 
of Amphioxus agreed very closely with that of Stieda, but he made 
a more detailed study of its histological character, and defined its 
limits more precisely. He concludes that the beginning of the 
spinal cord proper, in the absence of any outward mark of dis- 

Fig. 51. Sagittal section through the cerebral vesicle of Amphioxus. (After 

c.v. Cavity of cerebral vesicle, e. Eye-spot, g.c. Dorsal group of ganglion- 
cells (cf. Fig. 46). inf. Infundibular depression, l.o. Lobus olfactorius impar. 
tp. Tuberculum posterius. 

tinction from the brain-region, would lie at the point marked by 
the appearance of the first of the giant ganglion-cells, which he 
denotes by the letter A. (Cf. Fig. 48.) 

Quite recently the attempt has been made by Professor VON 
KUPFFER to determine in detail the delimitation of the cerebral 
vesicle of Amphioxus (Fig. 51). The slight outpushing of the 
wall of the vesicle towards the base of the olfactory pit has been 
mentioned in the text. It was discovered by LANGERHANS in 


1876, who called it the lobus olfactorius. Kupffer has succeeded 
in finding a similar structure in the embryos of other Vertebrates, 
notably in Acipenser sturio (the sturgeon). He calls it the lobus 
olfactorius impar, and shows that it indicates the point where the 
medullary tube remained for the longest and last time in direct 
connexion with the external ectoderm, precisely as is the case in 
Amphioxus. There is thus at least one fixed point common to 
the cerebral vesicle of Amphioxus and the brain of the craniate 
Vertebrates. But Kupffer has found another. While it is obvious 
that the anterior wall of the vesicle containing the pigment which 
constitutes the eye-spot is homologous with the primary optic tract 
(recessus opticus} of the higher Vertebrates, in which pigment is, 
in many cases, deposited in the embryo, Kupffer states that he 
is able to detect an infundibular depression in the floor of the 
cerebral vesicle of Amphioxus. Immediately behind this depres- 
sion there is a prominence in the wall of the vesicle, which Kupffer 
calls the tubercuhtm posterius. This point is also to be identified 
in the brains of the higher Vertebrates. 

The dorsal dilatation of the central canal, which Hatschek dis- 
covered and compared with the fourth ventricle of the vertebrate 
brain, whose roof is similarly membranous and not nervous (Fig. 
45), is certainly a very curious, and apparently constant, feature 
in young individuals, as I can affirm in confirmation of Hatschek. 
The conclusion come to by Hatschek, however, that the lobus 
olfactorius of Langerhans is the homologue of the infundibulum of 
the higher forms, would appear to be untenable in the light of 
Kupffer's researches. 

It is beyond the scope of this book to discuss the difficult 
problem of the origin of the paired eyes of the Vertebrates, but it 
may be pointed out that there is no difficulty in identifying a 
stage in the embryonic development of the optic tract in the 
Craniota corresponding to the permanent condition of things 
in Amphioxus. This fact was first demonstrated by WILHELM 
MULLER in 1874. On account of its position in front of and 
below the cerebral vesicle, it is doubtful whether the eye-spot of 
Amphioxus is homologous with the eye of the Ascidian tadpole. 
(See below.) 

10. (p. 94.) It is a significant fact that giant nerve-fibres appear 

NOTES. 103 

to be present in the spinal cord of all those Vertebrates whose tail 
serves as an important organ of locomotion. Thus, they occur in 
fishes, tailed Amphibia, in the tadpoles of tailless Amphibia, and, 
finally, they have been recently discovered by MAX KOPPEN in the 
caudal region of the spinal cord of the lizard. In the frog and higher 
forms they do not occur. From these considerations Koppen 
thinks that there is a causal relationship between the occurrence 
of giant-fibres in the spinal cord and the presence of a locomotor 
tail. The caudal locomotion, characterised by the rapid swaying 
motion of the tail, is not confined to the post-anal region in 
Amphioxus, but involves the whole body. 

Contrary to the observations of EISIG, both NANSEN and ROHDE 
are of opinion that the giant-fibres of Annelids (Polychaeta) have 
the same physiological significance for the central nervous system 
as those of Amphioxus. 

Some of the older authors mistook the giant nerve-fibres for 
capillary blood-vessels. As a matter of fact no blood-vessels 
traverse the central nervous system of Amphioxus. It may be 
added, also, that there are no medullated nerve-fibres. 

n. (p. 95.) Several suggestions have been made as to pos- 
sible representatives of the spinal ganglia of the dorsal roots of 
the Craniota in Amphioxus. 

Omitting earlier, and obviously erroneous, suggestions, ROHDE 
(1888) regarded the nuclei, which he found imbedded in the 
dorsal roots, as a collection of '' nervous nuclei," comparable to 
the spinal ganglia of the higher Vertebrates (Fig. 46). According 
to RETZIUS (1890) these nuclei are not of a nervous nature (prob- 
ably belong to supporting-cells), and he tentatively suggests that 
the spinal ganglia are represented by groups of bipolar ganglion- 
cells which occur inside the spinal cord at fairly regular intervals 
in two longitudinal rows, one on each side of middle line. The 
main process (axis-cylinder) of these cells divides in T-form, and 
one of the branches of the T passes into the dorsal root. (Cf. 

Fig. 5-) 

Finally, Hatschek (1892) finds the homologues of the spinal 
ganglia at the points where the dorsal nerves divide into ramits 
dorsalis and minus ventmlis. 



As an introduction to the study of embryology, and as 
an indispensable aid to a reasonable appreciation of the 
value of embryological facts, the life-history of Amphioxus 
provides an object which, for its capability of application 
to almost every branch of zoological discussion, is perhaps 
unrivalled. It is alike useful in a text-book of human em- 
bryology, and in one of invertebrate zoology. 

The reason for this obviously lies in the fact that in 
Amphioxus everything has its own definite line of de- 
marcation, all the fundamental structures of the body are 
laid down with schematic clearness, there are no massive 
agglomerations of cells forming complicated tissues, but all 
the organs are of simple epithelial origin and constitution. 

Whereas in many of the higher and lower animals the 
greatest difficulty is often experienced in deciding to which 
of the primary layers of the body this or that structure 
owes its origin, in Amphioxus there is no such difficulty. 
With these advantages it is, therefore, no wonder that 
Amphioxus should serve as a refuge to the perplexed 

It is not an exaggeration to say that the researches both 
of KOWALEVSKY and of HATSCHEK, on the development of 
Amphioxus, will always rank among the classics of embry- 
ological literature ; while it is a familiar fact that Kowa- 
levsky's earlier work (1867) on the development of the 



Ascidians and of Amphioxus marks a distinct epoch in 
the progress of the science of embryology. 

Fertilisation and Segmentation of the Ovum. 

The breeding-season of Amphioxus extends, in the Med- 
iterranean, from spring to autumn. 

The gonadic pouches become very much distended by 
the ripening of the ova and spermatozoa in the respective 
sexes,, and finally burst, discharging their contents into the 
atrial cavity, whence they 
reach the exterior through 
the atriopore. 1 At the time 
of complete sexual matu- 
rity the gonads become so 
large that the atrium is 
used up to its utmost 
capacity, and its walls be- 
come stretched to such an 
extent that the meta- 

pleural folds are flattened Fig. 52. -Unfertilised ovum 

oxus. Magnified about 750 diameters. 
Up against the sides of the (After LANGERHANS.) 

i i a. IUIK gianuie:=. /. romcie. 71. i\u- 

y* cleus (germinal vesicle), with nucleolus. 

The OVUm is Semi- $ P roto P' asm i c area, free from yolk gran- 
ules, surrounding the nucleus. 

opaque, contains granules 

of yolk equally distributed throughout its substance, and 
is surrounded by a cellular membrane known as \.}\Q follicle 
of the egg, and sometimes less accurately spoken of as 
the mtclline membrane (Fig. 52). 

Spawning, when it occurs, invariably takes place at sun- 
down, i.e. between five and seven o'clock in the evening, 
and never, so far as is known, at any other time. 



Ova and spermatozoa are discharged simultaneously by 
male and female individuals into the water, and fertilisa- 
tion is effected in the latter medium. 

The final stages in the maturation of the ovum of Am- 
phioxus are very imperfectly known, and the extrusion of 
the so-called polar bodies, preparatory to the process of fer- 
tilisation, has not been properly studied, only one such 

Fig. 53. Fertilised ovum of Amphioxus. Highly magnified. (From a 
drawing kindly lent by Professor E. B. WILSON.) 

d.c. Directive corpuscle or polar body. o. Ovum. f. Follicle. 

body having been observed, whereas from the analogy of 
all other sexually reproducing animals we should expect 
two polar bodies (directive corpuscles) to be given off be- 
fore the egg was fully mature. As soon as an ovum has 
been fecundated by the entrance of a spermatozoon, the 
vitelline membrane springs away from the body of the egg- 
cell, leaving a wide space between the latter and the 
former (Fig. 53). This expansion of the vitelline mem- 



brane is the first outward and visible sign of the accom- 
plishment of the process of fertilisation. 

About an hour later, that is to say, at about 8 P.M., - 
the egg becomes flattened at its two poles, and a depression 

Fig. 54. Division of ovum into the first two blastomeres. The polar body 
marks the animal pole. (After HATSCHEK.) 

appears at the animal pole, the latter being indicated by 
the polar body. The depression deepens and extends as a 
meridional furrow round the egg. Finally, the division of 
the egg into two halves or blastomeres, which remain at- 
tached to one another, is completed, and the first stage in 
the segmentation of the egg is accomplished (Fig. 54). 

As it is beyond the scope of 
this book to discuss the mechan- 
ics of cell-division, the descrip- 
tion of the segmentation stages 
will be very brief. 

The first meridional cleavage 
which divides the egg into two 
blastomeres is followed by an- 

nt-h^r nnf > Pt rJo-Kr nn crlpc tn it Fig ' 55- E 'g ht - ce11 sta g e seen 

right angle it, from the U p per (animal) pole . Four 

dividing each Of the tWO blastO- sma11 blastomeres (micromeres) lie 

upon four larger blastomeres fma- 
mereS again into tWO. In this cromeres). Radial type of cleavage. 

way the stage with four equal (After E. B. WILSON.) 
blastomeres in one plane is produced. Next follows an 
equatorial cleavage, by which eight blastomeres are pro- 
duced, the four upper cells at the animal pole being some- 



what smaller than the four lower cells at the vegetative 

pole, since the latter contain a greater quantity of the 

yolk-spherules (Fig. 55). 

The next cleavage giving rise to an embryo of sixteen 

cells is meridional. Then the eight upper and the eight 

lower cells become respectively 
divided by equatorial cleavages, 
and so the thirty-two cell stage 
is reached (Fig. 56). 

The embryo is now known 
as a blastnla, and consists of a 
mulberry-like mass of cells sur- 
rounding a central cavity called 

Fig. 56. Thirty-two cell stage, .... 

consisting of four rows of eight ceils, the segmentation-cavity or bias- 

each ranged around a central seg- fnrrp] 
mentation cavity (blastocoel). The 

polar body is still visible at the ani- From this point of the de- 
nial pole. (After HATSCHEK.) 

velopment the blastomeres go 

on dividing with more or less regularity, until the wall of 
the blastula consists of a great number of cells arranged 
in a single layer about the central cavity. 

The segmentation of the egg of Amphioxus, however, 
by no means follows the uniform and stereotyped plan 
that has been hitherto supposed. It has recently been 
discovered by Professor E. B. WILSON that Amphioxus 
presents an example of a polymorphic cleavage. Instead 
of following one type, it follows three types of cleavage ; 
namely, a radial type (Figs. 55 and 56), a bilateral type 
(Fig. 57), and a spiral type (Fig. 58). These three types 
of cleavage are reducible to a common basis, and are con- 
nected together by all possible intermediate gradations. 
Wilson points out that in the bilateral type of cleavage 
Amphioxus shows a close correspondence with the Ascid- 
ian embryo. 



The segmentation or cleavage of the ovum results in 
the formation of a spherical blastula, closed at all points, 

Fig. 57. Three stages in the segmentation of the ovum, according to the 
bilateral type. From the lower pole. (After E. B. WILSON.) 

A. Eight-cell stage. A, B, C, D. The four macromeres, above which are seen 
portions of the four micromeres. 

/-/. Plane of first cleavage, with respect to which the cells divide in such a way 
as to become arranged in a bilaterally symmetrical manner. 
//-//. Plane of second cleavage. 

B. Transition to the sixteen-cell stage. 

C. Sixteen-cell stage. The line in each cell indicates the direction in which the 
next division of the cell would take place. 

and consisting of some 256 cells surrounding a spacious 
cavity, the blastocoel. 

The stages of development lead- 
ing up to the blastula are known 
as the segmentation stages. At 
their completion, although, of 
course, cell-division continues to 
take place actively, yet other pro- 
cesses supervene which render the Fig- 58. Eight-cell stage 

. . from the upper pole, illlustrat- 

mere division of the individual cells mg the spiral type of cleavage. 

of minor morphological importance. < After E " R WlLSON -) 


The next phase of the development is known as the 
gastrulation of the embryo. The cells forming the lower 
or vegetative side of the blastula remain, throughout the 
segmentation stages, somewhat larger than the rest of the 



blastula-wall. This side now becomes flattened, as shown 
in Fig. 59 A. Next, the flattened side of the blastula 
becomes gradually tucked up or invaginated into the 

blastoccel (Fig. 59 B) until, 
finally, the segmentation 
cavity is completely obliter- 
ated, and the invaginated 
layer of cells becomes tightly 
fitted up against the outer 
layer (Fig. 59 Q. 

c 9% The embryo, now known 
as the gastrnla, is a double- 
layered sac, the cavity of 
which was produced by in- 
vagination, and is known as 
the primitive gastral cavity, 
or archenteron. This cavity 
is widely open to the ex- 
terior by the orifice of invagi- 
nation, or blastopore, which 
in German is designated by 
the expressive term Urmnnd. 
The two layers of cells which 
constitute the wall of the 
gastrula are the primitive 
germ-layers ; the outer layer 
is the primitive ectoderm, 
and the inner layer, sur- 
rounding the gastral cavity, 
is the primitive endoderm ; the two layers are continuous 
with one another round the margin of the blastopore. 

The blastopore is at first a very wide oval opening, 
but it soon becomes narrowed down to a small aperture 

o r 8 .a 

~ rt u 

Z .2 8 



by the continued deepening of the archenteric cavity 
(Fig. 60). 

It is now a well-established fact that all multicellular 
animals (Metazoa) pass through a gastrula-stage in the 
course of their development, although the form of the 
gastrula is often extremely modified and difficult to recog- 
nise. The significance of this 
fact, as was long since pointed 
out by Huxley, Haeckel, Lan- 
kester, and others, is very 
great when it is remembered 
that the embryonic character- 
istics of the gastrula are 
essentially identical with the 
permanent features of the 

Fig. 60. Optical longitudinal sec- 
Of the Coelen- tion of i ate r gastrula. Cilia (flagella) 


terata (Hydra, etc.). 

have been protuded from the ectoderm 
cells, and the embryo at this stage 
Returning tO the gastrula begins to rotate within the follicle. 

(After HATSCHEK.) 
of Amphioxus, in the course 

of the further differentiation which goes hand in hand 
with the progressive growth and development, we shall 
find that the primitive ectoderm gives rise to (i) the 
central nervous system and (2) the definitive ectoderm ; the 
primitive endoderm gives rise to (i) the mesodcrm, which 
is usually regarded as a third or intermediate germ-layer ; 
(2) the notochord ; and (3) the definitive endoderm, which 
forms the lining mucous epithelium of the alimentary 
canal ; finally, the primitive gastral cavity or archenteron 
will become subdivided into (i) the body-cavity or ccelom, 
and (2) the definitive gut or alimentary canal. 

The embryo shown in optical section in Fig. 60 repre- 
sents the stage reached at midnight of the first night of 
development. It will be noticed that one side is convex. 



while the opposite side is flattened ; this is an indication 
that dorso-ventral differentiation has taken place, since 
the flattened side marks the dorsum or back of the embryo, 
while the convex side is ventral. It may be seen further 
that the blastopore is inclined towards the dorsal side of 
the embryo. The dorsal inclination of the blastopore is 
eminently characteristic of the vertebrate gastrula from 
the Ascidians up to the highest 
craniate forms. In the Inverte- 
brates (Annelids, Molluscs, etc.) 
the blastopore acquires a ventral 

At the stage represented in Fig. 
60 the embryo commences to ro- 
tate within the vitelline membrane, 
each ectodermic cell being now 
provided with a vibratile cilium. Fig. 61. Elongated gas - 

,, . . trula. Optical longitudinal sec- 

The embryo next begins to elon- tion- The cilia are omitted 

gate, and the blastopore becomes from the ectoderm - (A fter 


still narrower (Fig. 61). 

A comparison of the accompanying figures will show 
that the narrowing of the blastopore is effected by the 
downward and backward growth of its dorsal border, 
while the ventral lip remains stationary. The dorsal ecto- 
derm, which is converted into the medullary plate, now 
shows indications of a shallow longitudinal groove. This 
is the beginning of the medullary groove which leads on 
to the formation of the central nervous system. 

* For a discussion of the phylogenetic relation of the blastopore or proto- 
stoma (Hatschek) to the mouth and anus, the following works should be 
consulted: ADAM SBDGWICK, On the Origin of Metameric Segmentation, etc., 
Quarterly Jour. Micro. Sc., XXIV., 1884, and by the same author, Notes on 
Elasmobranch Development, Ib. Vol. XXXIII., 1891-92. 

Finally, BERTHOLD HATSCHEK, Lehrbuch der Zoologie, Jena, 1888-91. 


Growth of Free-swimming Embryo. 

Between 4 and 5 A.M. in the first morning of develop- 
ment, i.e. at about the eighth hour, the embryo has reached 
the stage represented in Fig. 62, and it now bursts through 
the vitelline membrane and becomes free, swimming by 
means of its cilia at the surface of the sea, or aquarium. 

The fact that Amphioxus has a free-swimming, ciliated 
embryo is important as providing a general connecting 
link between the Vertebrates and the Invertebrates, since 




Fig. 62. Embryo of Amphioxus at the stage at which it ruptures the follicle 
and becomes free-swimming. 

A. Seen from above as a semi-opaque object. (After KOWALEVSKY.) 

B. Seen in sagittal (optical) section. (After HATSCHEK.) 

arc. Archenteron. m.p. Medullary plate. my.c. Myocoelomic pouches of 
archenteron. p.n.c. Posterior neurenteric canal. 

the possession of a ciliated ectoderm is very common 
among Invertebrate embryos, but entirely unknown among 
the craniate Vertebrates. 

The medullary plate is now being closed off from the 
outer surface. This is effected by the co-operation of two 
factors. The ectoderm which bounds the medullary plate 
laterally, grows over it, and simultaneously the ectoderm of 
the posterior or ventral lip of the blastopore grows for- 
ward over the medullary plate so as to shut in the blasto- 
pore from the exterior (Fig. 62 A and B}. The archenteric 


cavity therefore no longer opens by the blastopore to the 
exterior, but it communicates with the medullary tube. 
The blastopore has, in fact, become converted into the 
neurenteric canal, joining the canal of the central nervous 
system with the cavity of the alimentary system. This 
remarkable condition of things was first discovered by 
KOWALEVSKY, who also found it in the Ascidians and in 
a number of the higher Vertebrates. It has since been 
found to occur in all classes of Vertebrates, including 

Hitherto the body-wall of the embryo has consisted of 
only two primary germ-layers, ectoderm and cndodcrm. 
At the stage now under consideration, a third interme- 
diate layer, the mesoderm, has begun to put in its appear- 
ance. The mesoderm arises in the first instance as a 
series of paired lateral pouches of the archenteron. In 
Fig. 62 the first two or three archenteric pouches are 
distinctly visible. Before proceeding, however, to a more 
detailed account of the origin of the nervous system and 
of the mesoderm, we will trace briefly the changes in 
external appearance which the embryos undergo up to the 
time of the formation of the mouth. 

As the embryos are very transparent, the external 
appearance necessarily involves a good deal of the inter- 
nal structure. 

The period of embryonic development may be defined as 
commencing with the first cleavage of the ovum, and end- 
ing with the perforation of the mouth, thus comprising 
approximately the first thirty-six hours. During this 
period the embryo does not take up independent nourish- 
ment, but lives on the original food-yolk which was con- 
tained in the egg. 

During the first few hours of its pelagic or free-swim- 


ming existence, the embryo keeps rigidly to the surface 
of the water. 

After its escape from the vitelline membrane, it grows 
rapidly in length. Fresh archenteric pouches are added 
to those already formed, one after the other, in metameric 
order. The medullary plate (i.e. the fore-cast of the nerve- 
tube) becomes completely closed in beneath the superficial 
ectoderm except at its anterior extremity, where it remains 
open to the exterior in the mid-dorsal line by an aperture 
known as the ncuropore (Fig. 63 A, B, C). Finally, the 
notochord becomes differentiated from the primitive endo- 

According to Hatschek the number of mesodennic 
somites which arise as diverticula from the archenteron 
is fourteen pairs. Those which are subsequently added 
to these arise at the hinder end of the body by prolifera- 
tion from the cells which lie behind, and at the sides of 
the neurenteric canal, or in that region, so that they never 
appear as actual outgrowths from the archenteron. 2 

In Fig. 63 C the embryo has undergone some radical 
changes in form. Its body, previously cylindrical, has 
become laterally compressed, the ectoderm cells of the 
hinder end of the body have begun to elongate so as to 
form the rudiment of a provisional caudal fin, and the 
front end of the body has grown out into the shape of 
a snout. In connexion with the latter there are two 
remarkable structures which arise as a pair of outgrowths 
from the anterior region of the archenteron, and were first 
described by Hatschek as a pair of anterior intestinal 
diverticula. These we shall return to later. 

Near the front end of the alimentary canal a curious 
sac-like structure has appeared (Fig. 63 C}. It arose as 
a transverse groove in the floor of the gut in the region 


of the first myotome, extending from the right side under- 
neath to the left side of the body. (Cf. Figs. 63 A and 71.) 
The groove deepened, and its margins coalesced, and so it 

Fig. 63. Growth of the ciliated embryo of 
Amphioxus. (After HATSCHEK, slightly altered.) 

A. Stage, with nine pairs of myocoelomic pouches ; 
from left side. 

B. Same stage from dorsal side. 

C. Stage, with fifteen pairs of myotomes; from 
the right side. Vacuoles have appeared in cells of 

ch. Notochord. c.s.g. Club-shaped gland, g.s. 
Rudiment of first gill-slit, int. Intestine, l.a.d. Left 
head-cavity (left anterior intestinal diverticulum). 
my.c. Myocoelomic (archenteric) pouches, tip. Neu- 
ropore. n.t. Medullary tube. pg. Pigment granules 
in floor of medullary tube, p.n.c. Posterior neuren- 
teric canal, r.a.d. Right head-cavity (right anterior 
intestinal diverticulum). 

became constricted from the gut, and now forms a hollow 
sac closed at present at both ends. It is known as the club- 
shaped gland. Immediately behind it, in Fig. 63 C, is seen 



a shallow depression in the floor of the 
gut. This is the indication of the first 
gill-slit which becomes perforated at 
this point later. 

From this stage it is an easy tran- 
sition to the stage which marks the 
close of the embryonic and the com- 
mencement of the lan>al period of 

In the embryo shown in Fig. 64, the 
mouth appears as an oval aperture placed 
asymmetrically on the left side. At its 
first origin it is relatively much smaller 
than shown in the figure. A disc-shaped 
thickening of the ectoderm appears on 
the left side, in the region of the first 
myotome. The subjacent endoderm 
fuses with the thickening, and then the 
centre of the disc becomes perforated, 
and so the mouth is formed. 

The club-shaped gland has acquired 
an opening to the exterior immediately 
below the mouth, on the left side ; 
while the body of the gland lies on the 
right side. 

Behind the club-shaped gland on the 



Fig. 64. Stage in which the external apertures of 
the body, prasoral pit, mouth, first gill-slit, and anus 
have become perforated. Age about 36 hours. From 
the left side. (After HATSCHEK.) 

al. Alimentary canal, an. Anus. b.c. Body-cavity. Fig. 64. 

ch. Notochord. end. Endostyle. gl. Club-shaped gland, 
which has acquired an opening to the exterior on 

the left side below the mouth. g.s '. First primary gill-slit, m. Mouth, n.c. Xerve- 
tube ; the neurenteric canal has closed up, but the nerve-tube still curves round 
the hinder end of the notochord. np. Neuropore. p.o.c. Prasoral ccelom (right 
head-cavity), p.p. Praeoral pit (left head-cavity), t. Provisional caudal fin. 


right side is the first gill-slit, opening directly to the 
exterior. At the time of its actual perforation it lies 
near the mid-ventral line of the body, but as it increases 
in size it becomes shifted up to the right side. 

The neurenteric canal is closed up, and the nerve-tube 
ends blindly behind, being curved round the hinder end of 
the notochord. Immediately in front of and below the 
point where the neurenteric canal formerly existed, the 
amis has now made its appearance, approximately, if not 
exactly, in the mid-ventral line * (Fig. 64). 

We will now return to consider more closely the exact 
development of the mesodermic somites, the notochord, 
and the nerve-cord. 

Development of Central Nervous System. 

As in the craniate Vertebrates, so in Amphioxus the 
medullary plate arises as a median unpaired longitudinal 
specialised portion of the dorsal ectoderm. The way in 
which it becomes separated from the superficial ectoderm 
has already been indicated above, but it can best be 
studied in transverse sections. 

In the sections shown in Figs. 65 and 66, the separation 
of the medullary plate from the ectoderm, and its subse- 
quent conversion into a closed tube, is so clearly illus- 
trated, that further description is unnecessary. A unique 
feature in connexion with the formation of the central 
nervous system of Amphioxus is, that the medullary plate 
sinks below and becomes covered over by the superficial 
ectoderm before it takes on the form of a closed tube, so 
that for some time it exists as a half-canal open dorsally 

* According to Hatschek, the anus breaks through slightly to the left of 
the middle line. 
















> fl , 























































^2 rt 










W ^- 













H V " 




<^| * o S 
" | S ---'g g, 

, l . r^ O . 

_>, . 'g 

O uT "u ** O -H 

S > M ^ ^ 3 

S-|s .-^5^ 

^ o S ^ 1* 

-" M fl JH 4J p 

C CJ r-i -S- "^ -' O 

"" '? 8 c o ^ c 

3 K '~ N_^ C H 

Q " S 

T3 O " rt 
^ X -< ._ 

"" M "S "S 3 

2 g c o c 

i S aj 

C >-.!)*- r- 

* '3 ~; -fr >-! 

^5 oj 

i i T3 

.2r a> js ^ o 

fe ~ H ai <i 
3 I 5 




against the ectoderm. Later the dorsal margins of this 
half-canal meet and fuse in the middle line, and so 
produce the medullary tube * (Fig. 66). 

Origin of Mesoderm and Ccclom. 

In consequence of the flattening and incurving of the 
medullary plate, pressure is brought to bear on the 
dorsal wall of the archenteron, and the dorso-lateral bor- 
ders of the latter acquire the form of two longitudinal 
grooves (Figs. 65 A and B). It is from these grooves that 
the archenteric pouches are split off. The grooves deepen, 
and in doing so become divided up into a series of 
pouches. Eventually the pouches become shut off from 
the archenteron gradually from before backwards, and 
then appear as closed cavities on either side of the 
notochord, which has, in the meantime, been developing 
(Fig. 65 F). 

In the higher Vertebrates the mesoderm arises as two 
solid, lateral, longitudinal bands, which are split off from 
the primitive endoderm. These mesodermic bands are at 
first unsegmented, and might be taken to correspond with 
the longitudinal grooves of the archenteron of Amphioxus, 
as described above. Later, only the dorsal portion of the 
mesodermic bands undergoes segmentation, while the 
ventral portion, which becomes hollowed out to form 
the general body-cavity, is never segmented in the crani- 
ate Vertebrates. (Cf. Fig. 33.) In Amphioxus the whole 
of the mesoderm is contained in the archenteric pouches, 
and is, therefore, at first entirely segmented. 

As soon as the pouches -have lost their primitive con- 

* In the Ascidian embryo the formation of the medullary tube takes place 
after the manner typical of craniate Vertebrates (see below, IV.). 



.n r 


nexion with the archenteron, they commence to extend 
dorsally and ventrally between the ectoderm and the in- 
ternal organs (Fig. 66). Meanwhile the cells forming the 
inner or visceral wall of the pouch adjacent to the noto- 
chord elongate transversely and longitudinally, and begin 
to form the plate-like muscle-fibres of the myotome. The 
cells producing these fibres 
are arranged in such a way 
that each of them gives rise 
to a muscle-fibre extending 
from the anterior to the pos- 
terior limit of a myotome.* 
The muscles are at first 
closely approximated to the 
notochord and project freely 

into the Cavity of the pOUCh. Fi S- 66. Transverse section through 

the middle of the body of an embryo, 

The latter gradually grOWS with ten pairs of somites, to show the 
j i . -i , closure of medullary tube and the dorsal 

downwards, until it meets and ventra i extension of the mesodermic 
its fellow of the Other side ; somites. (After HATSCHEK.) 

al. Alimentary canal, c/i. Notochord, 
the tWO fuse together, and in the cells of which vacuoles have com- 

so the cavity is made con- ^S *?: 

tinUOUS from Side tO Side, the cells forming the inner wall of the 

somite, my.c. Myoccelomic cavity. 

below the intestine. 

When this occurs, the primarily single cavity of each 
archenteric pouch becomes divided into two portions ; 
namely, a dorsal portion, the somite proper or myoccel, 
and a ventral portion, the ccelom, by a transverse partition, 
which arises through a fusion between the parietal and 

* Already in the embryo shown in Fig. 63 C, and even at a somewhat ear- 
lier stage, the muscles are so far developed that the body can be bent and 
jerked. By the time the mouth has broken through, muscular locomotion 
effectually replaces the primitive ciliary locomotion, although the cilia persist 
to a late stage. 



-n c 

~s. t. tf 

visceral walls of the cavity, at about the level of the base 
of the notochord (Fig. 67). 

The dissepiments between the myotomes are formed 
from the contiguous walls of the successive pouches, but 
ventrally, in the region of the coelom, they break down, 
so that the latter then becdmes a continuous unseg- 

mented cavity. On account 
of the fact that the archen- 
teric pouches give rise both 
to the cavity of the somites 
(myoccel} and to the general 
body-cavity (coelom proper 
or splancJinoccel}, they are 
often spoken of as the myo- 
ccelomic pouches. The cav- 
ity of the original archen- 

Fig. 67. Scheme of a transverse teric pouches is known as 

section through the body of a larva t , primitive ffflnni the 

with five gill-slits, to show the division "* ;/ ' LU 

between myoccel and splanchnoccel. epithelial walls of which 
(After HATSCHEK.) 

n.c. Spinal cord (medullary tube). Constitute the mcsoderm. 

^Notochord. l.m. Muscles. myMyo- Ag d iff erent iation and OF- 

coel. sf. Rudiment of sclerotome. 

al. Alimentary canal, s.i.v. Sub-intestinal ganogeny proceed, the mCSO- 

vein. sp. Splanchnoccel. 

derm gives rise to (i) the 

musculature, (2) the connective tissue, (3) the blood-vessels, 
(4) the reproductive organs, (5) the ccelomic epithelium or 
lining of body-cavity, also called the peritoneum, and 
(6) the excretory tubules. The development of the last- 
named structures has, however, not yet been worked out 
in Amphioxus. 

The parietal layer of the mesoderm applies itself closely 
against the ectoderm, and gives rise to the cutis of the 

The connective tissue-sheath of the notochord and 



nerve-cord, together with the internal sheath or fascia 
of the muscles of the myotome, arises from the walls of 
a pouch-like diverticulum of myoccel which grows up be- 
tween the muscles and the notochord and nerve-cord. (Cf. 
Figs. 67 and 68.) The myoccel also grows downwards 
between the somatic layer of the peritoneum and the ecto- 
derm (Fig. 68). According 
to Hatschek the dorsal and 
ventral fin-spaces are also 
derived from the myoccel. 3 

The diverticulum of the 
myoccel which has just been 
described is known as the 
sclerotome, since it gives rise 
to the fibrous sheath of the 
notochord and nerve-cord, 
which (i.e. the sheath) in 
most of the higher forms 
becomes replaced by carti- 
lage, and finally by bone. 
In the craniate Vertebrates Fig 68 _ Scheme of a transverse 

the SclerotOme arises as a section through region between atriopore 

and anus, of a young Amphioxus shortly 
Solid proliferation of cells after the metamorphosis. (After HAT- 

from the visceral wall at the ;> porsal fin . space _ ^ MyocoeU 

base of the SOmite. This 5C - Sclerotome. ao. Aorta, al. Intestine. 

i.m. Intercoelic membrane, s.i.v. Sub-in- 
solid proliferation is Ull- testinal vein. sp. Splanchnocoel. v.f.c. 

doubtedly a modification of Ventral fin ' space - 
a hollow diverticulum, involving, as it does, only the 
visceral wall of the somite, precisely as we find it in 
Amphioxus. 4 (Cf. Fig. 33.) 

On their outer surface the muscles of the myotomes are 
not provided with a sheath of connective tissue (fascia), 
standing, in this respect, in contrast to the condition 
which obtains in the Craniota. 


Origin of the Notochord. 

The notochord is formed from the endodermic cells 
which lie between the mesodermic pouches and constitute 
the dorsal wall of the archenteron. The dorsal wall of 
the archenteron at an early stage becomes converted into 
a shallow longitudinal groove whose concavity is turned 
towards the archenteric cavity (Fig. 65 D). This groove 
gradually deepens (Fig. 65 E], and eventually its walls 
become closely appressed to one another so as to obliter- 
ate the lumen (Fig. 65 F). Finally the adjoining cells of 
the archenteric wall grow across the gap occasioned by 
the formation of the notochord, and joining together, shut 
off the latter from any participation in the enteric wall 
(Fig. 66). In this way is the notochord separated from 
the endoderm gradually from before backwards. Poste- 
riorly it remains for a considerable time fused with the 
endoderm at the point where the anterior wall of the neu- 
renteric canal becomes continuous with the dorsal wall 
of the archenteron. 

We have indicated above that the differentiation of the 
notochord takes place from before backwards. At its 
anterior extremity a very noteworthy exception to this 
rule is presented. In the region of the first myotome 
the notochord retains an open communication with the 
archenteron after its lumen has already been obliterated 
behind this point. Moreover, in the embryo, with eight 
pairs of myocoelomic pouches (Fig. 68 bis), the front end 
of the notochord lies some distance behind the front end 
of the body, while the anterior portion of the archenteron 
extends beyond the notochord. Eventually the notochord 
is continued to the front end of the body by becoming 
constricted off from the dorsal wall of the anterior sec- 




tion of the archenteron in the usual way. This retarded 
growth of the notochord anteriorly indicates that its exten- 
sion to the tip of the snout is a secondary phenomenon. 
Ancestrally we are bound to assume it did not extend so 
far forwards. The forward 
extension of the notochord 
is, as noted above, obviously 
useful to Amphioxus in ren- 
dering its pointed snout 
sufficiently resistant to en- 
able it to burrow in the 
sand. When it wants to 
bury itself in the sand, it 
has not to take pains to dig 
a hole, but darts in in the 
fraction of a second. 

The histological differen- 
tiation of the notochord 
commences soon after the 
sides of the chordal groove wi ,f * 

have Come together SO as tO primary relations of the anterior end of 

the notochord. From above. (After 

obliterate the lumen. 1 he HATSCHEK.) 

cells composing the noto- ** ? s ? c i rdal portion J ar t cl T 

teron, which becomes converted into the 

Chord are, at the first ap- head-cavities. ./. Neuropore. ch. Noto- 

. chord; over which lies the neural tube. 
proximation Of the walls OI my _ Myocoelomic pouches, ne. Neuren- 

the groove, placed end to teri ^ T c * naL T 

N.B. In this and other figures of 

did, but SOOn begin tO inter- Amphioxus embryos here reproduced 
. after Hatschek, the so-called mesoder- 

lace with one another across mic pole cells have been emitted in 

the middle line (Fig. 65 F\ accordance with the observations of 

and finally each cell comes 

to occupy the whole width of the notochord (Fig. 66). 

Meanwhile vacuoles begin to appear in the cells (Fig. 66). 

The vacuolisation of its component cells is an extremely 


characteristic feature of the notochordal tissue throughout 
the group of the Vertebrates. It is carried on to such an 

extent in Amphioxus as to 
obscure the original cellular 
structure of the notochord. 
The cells anastomose with 
one another in the longitu- 
dinal direction, and so pro- 
duce a reticulum the meshes 
Fig. 69. Median sagittal section of of which represent the vacu- 

notochord of a young Amphioxus of o]es whose firgt { j 
8 mm., to show the vacuolar character 

of the notochordal tissue and the dis- shown in Fig. 66. Most of 
placement of the nuclei to the dorsal and . 

ventral borders. (After LWOFF.) the nuclei become eventually 

displaced from the centre of 

the notochord, and are, in the adult, almost exclusively 
confined to its dorsal and ventral aspects (Fig. 69). 

The Prcsoral "Head-cavities" of Amphioxus. 

Before leaving the embryonic period of the development 
it is necessary to consider the origin and fate of what may 
be called the head-cavities of Amphioxus as made known 
to us by the work of Hatschek. 

They arise symmetrically as a pair of diverticula from 
the anterior portion of the archenteron, which lies at first 
partly in front of the notochord (Fig. 68 bis) and completely 
in front of the myoccelomic pouches (Fig. 70). 

They begin to appear at the stage in which some eight 
pairs of pouches are already present. Their origin there- 
fore, in point of time and the subsequent modifications 
which they undergo, show that they do not belong to the 
metameric series of the mesodermic pouches, but are 
structures sui generis. 





The archenteron extends at first to the front end of the 
body. Its anterior portion, after the formation of several 
mesoblastic somites, becomes marked off from the hinder 
region by a slight constriction, which gradually becomes 
deeper and deeper (Fig. 70), until finally the whole of this 
portion of the archenteron is divided into two separate 
sacs, which eventually lose 
all connexion with the ar- 
chenteron (Fig. 71). Theali- 
mentary canal now no longer 
reaches to the anterior ex- 
tremity of the body. Very 
soon after their separation 
from the archenteron these 
sacs enter upon a series of 
changes by which their origi- 
nally symmetrical disposi- 
tion is entirely destroyed. 

Already in Fig. 71 it can 
be noticed that the right 
sac is becoming larger than Fi g. 70 . - Embryo, with nine pairs of 

the left, and the epithelium primitive somites seen in optical section 

from the ventral surface, to show the 
lining its walls is losing its origin of the head-cavites. (After HAT- 

C(~"l-TT?TC \ 

original cubical character, rM ' d . Right head-cavity. i.a.d. Left 

the inner ends of the Cells head-cavity, my.c :. Myocoelomic pouches 

(pnmtive somites), arc. Archenteron. 

are rounding off, and in fact 

it is being converted from a cubical to a flattened pavement 
epithelium (Figs. 63 C and 64). The left sac, on the con- 
trary, retains its original form and dimensions for a long- 
time. During the asymmetrical changes affecting the two 
sacs, which take place coincidently with the formation of 
the snout, the left one comes to lie transversely below the 
notochord, while the right sac becomes greatly enlarged 






and constitutes the cavity of the snout lying below the 
notochord (Fig. 63 Q. 

Shortly after the breaking through of the mouth the 
left sac acquires an opening to the exterior on the left side 
of the body (Fig. 64). The right sac becomes the /rawer/ 
body-cavity or ccelom of the "head," while the left sac is 
known as the prceoral pit. It is necessary to emphasise 
the fact that these two structures which are so different 

in their fully formed con- 
dition are at first perfectly 
similar and symmetrical and 
form a pair of "head-cavi- 
ties." Ultimately, as we 
have seen, only one of them 
H actually persists as a head- 
Fig. 71. - Anterior portion of em- cavity ; namely, the right one. 

bryo, with thirteen primitive somites, The ent j re conversion o f 
from the ventral side m optical section. 

(After HATSCHEK.) the left sac into the praeoral 

r.a.d. and l.a.d. Right and left head- . . 

cavities. c. s .g. Rudiment of club-shaped pit is probably to be regarded 
gland - as a secondary or cenoge- 

netic phenomenon, but the acquirement of an opening to 
the exterior is probably not secondary, since a similar 
opening (the proboscis-pore) occurs in Balanoglossus. 

In addition to the above-described peculiarities which 
sufficiently distinguish the head-cavities from the myocoe- 
lomic pouches, must be mentioned the fact that at no point 
of their epithelial walls are muscles developed. 

It is probable that the praeoral head-cavities of Amphi- 
oxus are homologous with the prcemandibnlar cavities of 
the higher Vertebrates, from the walls of which the greater 
number of the eye-muscles are developed.* This view is 

* This is also the opinion of Kupffer. Singularly enough van Wijhe has 
advanced the view that only the right head-cavity of Amphioxus is to be 



strongly confirmed by the mode of development of the 
praemandibular cavities in the Cyclostomes. 

In these fishes, VON KUPF- 
FER has shown that they 
actually appear in the form 
of a pair of diverticula from 
the anterior extremity of 
the archenteron (Fig. 72). 
If a comparison be made 
between Figs. 70 and 72, it 
will be at once manifest how 
close the correspondence is 

Fig. 72. Horizontal projection of 

between the mode Of de- pharynx and prasoral endodermic exten- 
c , , . sion of a young Ammocastes planeri of 

VClopment of the head-cavi- ^ mm _ ^constructed from a series of 
ties in AmphioxUS and in transverse sections. (After KUPFFER.) 

p.e. Prasoral endodermic extension 

AmmOCOeteS. In the Se- (prseoraleEndodermtasche). pm. and m. 
, ., . Prcemandibular and mandibular portions 

ltV 1S of head-cavities, ph. Cavity of pharynx. 
/, 2, j. First three pairs of gill-pouches. 

N.B. Kupffer considers it probable 
that the mandibular as well as the prae- 
mandibular cavities arise from the single 
pair of endodermic diverticula. In the 
course of the following pages I have 
referred chiefly to the prasmandibular 
In Fi~ 64. there is tO be cavities alone so as to avoid complica- 

noticed a vertically placed 

structure lying in front of and contiguous with the club- 
shaped gland. It is a tract of very high cylindrical cells 
forming part of the right wall of the alimentary canal in 

homologised with the prcemandibular cavity (see below, V.) . Kupffer regards 
the pnemandibular and mandibular head-cavities as rudimentary or meta- 
morphosed gill-pouches. This deduction is entirely foreign to the standpoint 
which I have adopted. The conclusion may seem plausible from the con- 
ditions observed in Acipenser alone; but when these are regarded from a 
comparative point of view, the deduction is, to my mind, unjustified. It should 
be added that Kupffer has shown that the head-cavities (prsemandibular and 
mandibular) of Acipenser also arise as endodermic pouches. 

hardly less striking. 5 

Endo style and Pigment 


this region. (Cf. Figs. 65 G and 75.) I have shown that 
this epithelial tract is the rudiment of the endostyle (vide 

It is a curious fact that the first trace of pigment to 
appear in the nerve-tube is not the eye-spot, but that at a 
constant point in the region of the fifth somite a black 
pigment-spot is deposited in a cell in the ventral wall of 
the medullary tube. This is followed by another smaller 
pigment granule slightly posterior to the first (Fig. 63 C). 
The eye-spot appears at the end of the embryonic period. 


Formation of Primary Gill-slits, etc. 

With the establishment of the definite relations oi the 
head-cavities, the mouth, club-shaped gland, first gill-slit, 
and anus, the embryo enters upon the larval phase of the 

It is no longer, or only very rarely, to be taken from 
the surface of the sea, but descends to a depth of several 
fathoms. When kept in aquaria, the larvae can often be 
observed to be suspended vertically, and apparently quite 
motionless in the water. This suspension is, no doubt, 
effected by the movement of the long cilia, or flagella, 
with which the ectoderm is provided, each cell possessing 
one flagellum. 6 

The principal changes which take place during the early 
stages of this phase of the development are the addition of 
new myotomes, the formation of new gill-slits, in meta- 
meric order, in an unpaired series on the right side of the 
larva, to the number of from twelve to fifteen, or even 
sixteen (the more usual number being fourteen), and the 
origin of the atrial cavity. 


Each gill-slit breaks through in, or slightly to the right 
of, the mid-ventral line, and then grows well up on the 
right side of the body. A larva with three gill-slits and 
the indication of a fourth is represented in Fig. 73. The 
originally circular mouth has grown to a much larger size, 
and extends on the left side anterior to the endostylar 

- 73- Larva of Amphioxus, with three gill-slits and the rudiment of a 
fourth ; from the left side. (After LANKESTER and WlLLEY.) 

/./. Praeoral pit. end. Endostyle lying on right side, seen through the wide 
lateral mouth, gl. Position of external aperture of club-shaped gland. p.s '. First 
primary gill-slit, an. Anus. 

N.B. Actual length of larva, nearly iV 2 mm. 

tract (which is on the right wall of the pharynx) and 
posterior to the first gill-slit. The oral opening later 
attains to relatively gigantic dimensions, and forms one 
of the most striking features of the larva. 

The anus is now displaced from its original ventral 
position to the left side in consequence of the increased 
development of the provisional caudal fin. The latter 
consists of elongated ectodermal cells, in which a certain 
amount of brown pigment is deposited. Later, when 
the dermal expansion, which has been described above as 
the definitive caudal fin, begins to grow out, it pushes the 
cells composing the provisional fin before it, so that they 
form a fringe round its border. Eventually the provisional 
fin disappears entirely. 

The gill-slits now go on adding to their number, one 
after the other, until the larva reaches the stage shown in 
Fig. 74. In this larva there are fourteen primary unpaired 
gill-slits, lying, for the most part, on the right side of the 



pharynx, although the more posterior slits bend under the 
pharynx, while the most posterior have a median ventral 

In front the gill-slits still open directly to the exterior, 
but the right metapleural fold is seen to be hanging over 
the tops of them ; while the hinder slits now open into 
the partially formed atrium, which has already closed in 

Fig. 74. Anterior portion of larva, with fourteen primary gill-slits and rudi- 
ments of the secondary gill-slits ; viewed as a transparent object from the right 
side. (After WILLEY.) 

s.o. Sense-organ of prasoral pit (groove of Hatschek). e. Endostyle. gl. In- 
ternal opening of club-shaped gland, s.s. Rudiments of secondary gill-slits. p.s iA , 
/.J 14 . Thirteenth and fourteenth primary gill-slits. The lower margin of the 
mouth is seen through the anterior gill-slits. 

Total length of larva, nearly 3% mm. 

posteriorly, as described above. The larva is remarkably 
transparent, so that its internal organisation can be seen 
as clearly as possible through the outer body-wall. 

The long axis of the primary gill-slits is approximately 
at right angles to the long axis of the body. They are 
not more numerous than the myotomes in the correspond- 
ing region of the body, so that the branchiomery at this 
stage coincides with the muscular metamery. In Fig. 73 
the first gill-slit was somewhat larger than the second and 
third. At about that stage, however, its further growth 
became arrested, and now it is seen to be considerably 
smaller than those which immediately follow it. 

In addition to its external opening on the left side, be- 






Fig- 75- Transverse sections through the region of the mouth of larvce of 
Amphioxus, to show the endostyle and the external and internal openings of club- 
shaped gland. (After LANKRSTER and WILLEY.) 

A. Section passing through the anterior corner of the mouth of a larva, with 
eleven gill-slits. 

B. Section passing through the middle of the mouth of a larva, with twelve 


al. Pharyngeal cavity, b.c. Ccelom (splanchnoccel) . br.e. Branchial epithelium. 
e.a. Branchial artery, end. Endostyle. ex.o. External opening of club-shaped 
gland. f.c. Dorsal fin-space, gl. Lower portion of club-shaped gland, g.s '. First 
gill-slit. i.i. Interccelic membrane, in.o. Internal opening of club-shaped gland. 
La.. Left aorta; there is no corresponding right aorta in the larva, tii. Mouth. 
r.m. Rudiment of right metapleur; a mere ectodermic thickening in A ; a solid 
thickening of the cutis in B, in which two of the original enlarged ectoderm cells 
have become imbedded. These curious cells occur over a long stretch of the 
metapleural folds during this phase of the development, disappearing eventually. 

In B, the left metapleur is indicated by an ectodermic thickening immediately 
below the gill-slit, x. So-called nephridium of Hatschek, 



low the mouth (see Fig. 64), the club-shaped gland has 
now acquired an opening at its upper extremity, on the 
right side, into the pharynx. 7 The gland lies, as usual, 
behind, and closely approximated to, the endostylar tract, 
which is bent forwards on itself at its upper end (Figs. 75 
A and B\ 

Pigment-spots are present in great numbers at the base 
of the neural canal. The pigment is deposited in special 


Jt,.C 1 


Fig. 76. Transverse sections through the region of the prasoral pit. (After 

A. Through a larva, with twelve gill-slits and no atrium. 

B. Through a larva, in which the atrium was closed in over all the gill-slits 
except the first two. (Cf. Fig. 38 C.) 

a.r.m. Anterior median portion of right metapleur. p.o.c. Prasoral body-cavity 
(right head-cavity) ; this cavity becomes much reduced after the metamorphosis, 
and is largely filled up by gelatinous tissue, p.p. Praeoral pit. s.o. Sense-organ 
of prceoral pit (groove of Hatschek). l.o.h. Rudiment of left half of oral hood. 
my' . Sclerotome (diverticulum of myoccel my). Other letters as above. 

Section B is taken through a plane slightly posterior to section A. 


cells, the pigment-cells, which arise as modified epithelial 
cells of the central canal. These cells send out several 
branching processes, which lose themselves in the fibrous 
tract of the spinal cord. 

Already in the youngest larva namely, that shown in 
Fig. 64 the praeoral pit had become subdivided into two 
portions, which, however, retained a free communication 
with one another. 

In the course of the changes which the left head-cavity 
had to undergo in its conversion into the praeoral pit it 
had come to lie transversely below the notochord. Sub- 
sequently it extended itself, in the form of an offshoot, 
dorsally to the right of the base of the notochord. 

This offshoot from the prasoral pit appears to serve as a 
special sense-organ lying ultimately, as mentioned above, 
in the roof of the oral hood, whose function is possibly to 
test the water as it enters the mouth (Figs. 76 A and B, 
and Fig. 74, etc.). 

Formation of Secondary Gill-slits. 

Above the primary gill-slits in Fig. 74, and like them, on 
the right side of the body, is to be observed a longitudinal 
ridge provided with a series of nodal enlargements which 
alternate with the primary gill-openings, the first of them 
lying above and between the third and fourth primary slits. 
Each of these enlargements represents a thickening in the 
wall of the pharynx, which has undergone fusion with the 
body-wall beneath the right metapleural fold, in the angle 
formed by the latter with the body-wall. 

These metameric fusions of the pharyngeal wall with the 
body-wall are the forecast of a second row of gill-slits, whose 
relation to the primary row will become clear as we pro- 


ceed. With their appearance, the larva enters upon that 
phase of its development which has been called the later 
larval period. It is the period of the metamorphosis of 
the larva, during which the pronounced asymmetrical 
arrangement of the parts is exchanged for the partial, but 
not absolute, symmetry which we have noted in the adult. 
The metamorphosis, therefore, consists largely in the sym- 
metrisation of the larva. 

The simultaneous appearance of the six nodal thicken- 
ings in the exact position, shown in Fig. 74, is very 
constant. Shortly afterwards a minute perforation appears 
in the centre of each thickening almost simultaneously, 
except in the case of the first, which usually becomes 
perforated rather later than the others. The originally 
small circular openings of the secondary gill-clefts gradually 
increase in size and become oval in shape, their long axes 
being parallel to the long axis of the body, instead of at 
right angles to it as in the case of the primary slits. 

Next, the upper borders of the secondary slits begin to 
flatten, and later to show signs of curving downwards. 
The changes in shape, which affect the secondary slits at 
the stages now under consideration, may be expressed by 
saying that they are at first shaped like a biconvex lens, 
then like a plano-convex lens with the flat surface directed 
upwards and the convex surface downwards, and finally 
like a concavo-convex lens with the concavity directed 
upwards (Fig. 77). 

During these changes, which do not take place in all the 
secondary slits at the same time, the last one especially 
retaining for a long time its primitive shape, the walls of 
the successive slits become sharply rounded off and distinct 
from one another, and a new perforation makes its appear- 
ance in front, above, and between the second and third 

LAR VAL DE I 'EL OPME.\ ' 7 '. 


primary slits. This new slit constitutes the definitive first 
slit of the secondary series (Fig. 77). 

The larva shown in Fig. 77 presents a very different 
aspect from that shown in Fig. 74; the transition from one 
stage to the other is, of course, gradual, and all intermediate 
steps can be observed. In the stage which we are now 
considering (Fig. 77), the atrial cavity has become com- 
pletely closed up in front, so that now none of the gill-slits 
open directly to the exterior. 

None of the primary slits now lie entirely on the right 
side, but they have become bent under the pharynx, and 

Fig. 77. Anterior portion of larva, in which the secondary slits have become 
perforated, and the primary slits are passing round to the left side. From the right 
side. (After WlLLEY.) 

i.o. Sense-organ of preeoral pit. v. Right half of velum, e. Endostyle, grow- 
ing beyond the club-shaped gland gl. p.s . First primary slit, much reduced in 
size. s.s ' . First secondary slit. /.J 12 . Twelfth primary slit, behind which is to be 
seen a vestige of the thirteenth slit. 

thus extend round to the left side. This bodily migration 
of the primary slits from one side to the other occurs in 
correlation with the increase in size of the secondary slits, 
which, as they continue to grow, push, as it were, the 
primary slits before them, and so cause the latter to bend 
under the pharynx in the way described. The peculiar 
growth by which the primary gill-slits are gradually carried 
from the right to the left side, may be described as a trans- 
verse or rotatory growth affecting the pharynx in toto in 


the region of the secondary slits. Such of the primary 
slits as occur behind this region are not affected by the 
rotatory method of growth, and retain their original position 
in the mid-ventral line of the pharynx. 

It is to be noted also that there are only twelve primary 
gill-slits at this stage. Assuming that in the particular 
larva here figured there were originally fourteen primary 
slits, the fourteenth has closed up and vanished without 
leaving a trace, while a vestige of the thirteenth can still 
be recognised. The actual process involved in the closure 
and disappearance of a certain number of the primary slits 
can, as we shall see, be readily observed in the living larva. 

Clnb-sJiapcd Gland and Endostyle. 

The internal aperture of the club-shaped gland into the 
pharynx is exceptionally plain at this stage, and its refring- 
ent walls and relatively large size give it a curiously slit- 
like appearance. We shall find that the gland subsequently 
atrophies, but the most persistent part of it that is to say, 
the last part of it to disappear --is precisely the internal 
opening with its refringent border. 

The endostyle, whose primary position, as we have seen, 
was immediately in front of the club-shaped gland, now 
presents a remarkable condition. It has begun to grow 
backwards and downwards, being probably pulled down, 
so to speak, by the general rotatory growth of which we 
have spoken above ; and so the club-shaped gland no 
longer lies behind it, but upon it. The gland itself being 
disconnected with the wall of the pharynx, except at its 
upper end where it opens into the latter, is not affected 
by the complicated changes to which the pharyngeal wall, 
including gill-slits, mouth, and endostyle, is subjected, so 



that it forms a convenient pnnctnm fixnm with relation to 
which the growth of neighbouring structures, particularly 
that of the endostyle, can be determined. 

The upper and lower limbs of the endostyle are inclined 
to one another at an acute angle, and may be said to form 
two unequal sides of a triangle, the apex of which is 
directed backwards between the rows of secondary and 
the primary gill-clefts (Fig. 77). 

Between the two rows of slits on the right side of the 
body there is a blood-vessel, representing the anterior 
continuation of the sub-intestinal vessel, which ends blindly 
in front above the first primary slit. This is the future 
ventral branchial artery, with which we are already ac- 
quainted. When its final situation in the mid-ventral line 
below the endostyle is remembered, its position in the 
larva high up on the right side, as in Fig. 74, will appear 
very striking. 

Continued Migration of Primary Gill-slits. 

The secondary slits now go on growing in size, and the 
primary slits gradually tend to disappear entirely from the 
right side until, as in Fig. 78, only the original upper por- 



Fig. 78. Anterior portion of larva from right side, to show the backward 
growth of the endostyle between the primary and secondary gill-slits. (After 

s.o. Sense-organ of praeoral pit. p.s' ' . First primary slit. in. Internal opening 
of club-shaped gland, e. Endostyle. p.b. Peripharyngeal ciliated band. 


tions of them are visible from this side. In some of the 
secondary slits the dorsal margin, which had previously 
begun to curve downwards, has now reached the ventral 
margin and fused with it (Fig. 78, third secondary slit). 
In this way is the tongue-bar formed, and the primitively 
simple gill-opening is divided into two distinct halves. 
The formation of the tongue-bars occurs in the secondary 
slits considerably in advance of the primary, both actually 
and relatively, since the latter have existed all through the 
earlier period of the larval development without a trace of 

Peripharyngeal Bands. 

The endostyle has now grown a long distance behind 
the club-shaped gland, and extends backwards between 
the two rows of slits as far as the middle of the second 
secondary slit. From the anterior part of the upper half 
of the endostyle, which is now nearly equal in length to 
the lower half, arises an epithelial tract in the wall of the 
pharynx, which appears in the form of a band of ciliated 
cells, and proceeds backwards below the notochord to the 
end of the pharynx. A corresponding ciliated band occurs 
in the left wall of the pharynx, proceeding from a similar 
point in the lower limb of the endostyle. In their course 
below the notochord the two bands take part in forming 
the hyperpharyngeal (dorsal) groove of the pharynx. 

Atrophy of First Primary Gill-slit and Club-shaped 

Gland, etc. 

We have already seen indications of a reduction in the 
size of the first primary slit. This reduction has advanced 
considerably in the stage we are now describing (Fig. 78), 
where the slit in question is only recognisable in side 
view as a small groove. 


The next stage to be considered is characterised above 
all by the simultaneous atrophy, closure, and disappearance 
of the club-shaped gland, and the first primary gill-slit 
(Fig. 79). At this stage the increase in size of the 
secondary slits has progressed to such an extent that the 
primary slits have been displaced entirely from their 
original position, and are no longer to be seen from the 


Fig. 79. Anterior portion of larva from right side after the disappearance of 
the club-shaped gland. (After WlLLEY.) 

s.o. Sense-organ, e. Endostyle. p.b. Peripharyngeal band. s.s'. First secondary 


right side, except in the case of the hindermost slits of 
the series, which remain, as mentioned above, in a median 
ventral position until their disappearance. 

A larva seen from below, so as to show the relative 
positions of the gill-slits and endostyle, etc., at this stage, 
is represented in Fig. 80. 

It is obvious, from what has been said above, that in the 
passage of the primary slits from their original position on 
the right side of the body to their final position on the left 
side, their dorsal and ventral margins are reversed. What 
was at first the dorsal edge of a primary slit becomes its 
ventral edge, and vice versa. In other words, what is 
actually the dorsal border of the primary slits in Fig. 74 
is morphologically the ventral border ; and conversely, what 
is actually the latter is morphologically the former ; and it is 



from the latter, towards the completion of the rotatory 
growth, which carries the slits from one side to the other, 
that the tongue-bars arise (Fig. 80). 

The vertical and longitudinal axes of most of the slits, 
both primary and secondary, are now almost equal, but 
the original difference in this respect, which we noted 
above, is still to be observed in the case of the foremost 
and hindmost slits of the two series. (Cf. Fig. 80, s.s 1 
and p.s 2 , and s.s s and p.s.) The first primary slit has 

e s.S' v ch 

p.s" p.s> 

Fig. 80. Anterior portion of larva of same age as in Fig. 79, seen from the 
ventral surface. The pharynx is flattened out. (After WlLLEY.) 

ch. Notochord. in. Entrance to mouth, v. Velum, p.s 1 . Vestige of first 
primary slit. f.s' 2 . Secondary primary slit. /.jio. Tenth primary slit. p.s^-. Ves- 
tige of twelfth primary slit. s.s 1 . First secondary slit. e. Endostyle. s.s. Eighth 
secondary slit. a. Atrium, pressed aside. 

now completely closed up, and its former existence is 
barely indicated by a loose granular appearance at the 
place it formerly occupied. 

The alternation of the gill-slits of the two series comes 
out very clearly in Fig. 80. In most of the secondary 
slits the formation of the tongue-bars is completed ; but 
not so in any of the primary slits, where it is only be- 

There are now eight secondary slits, an additional one 
having been added behind, alternating with the ninth and 
tenth primary slits. Usually the formation of secondary 
slits stops at this point, no more being formed until the 


number of primary slits is reduced to the same number ; 
namely, eight. 

Since it is usual for the primary slits to break through 
in the first instance to the number of fourteen, no less 
than six of them must close up and disappear before the 
stage with only eight gill-slits on each side of the body is 
arrived at. The six slits which are to close include the 
first and the five posterior primary slits. In the larva 
shown in Fig. So, the tenth and eleventh primary slits 
would have to close at a later stage ; the twelfth is on the 
point of closure, and its walls present the characteristic 
coarsely granular appearance spoken of above, while the 
thirteenth and fourteenth slits have entirely vanished. 

In addition to the fact of the closure of these primary 
slits, it is important also to emphasise the fact that they 
disappear without leaving a trace behind. In the higher 
Vertebrates there are a number of structures not only di- 
rectly connected at some stage of development with the 
pharyngeal wall, but also at some distance removed from 
it, which various morphologists have interpreted as the 
remnants of ancestral gill-clefts, without sufficiently con- 
sidering the question whether gill-clefts were in the habit 
of leaving their mark behind them. 8 In Amphioxus, at 
all events, they do not. 

The Adjustment of the Mouth, etc. 

While the gill-slits have been adjusting themselves to 
their definitive positions, the mouth has also been sub- 
jected to a peculiar kind of growth, which results in its 
bending round the front end of the pharyngeal wall, and 
ultimately assuming an anterior and median position, as 
we find it in the adult. 



In Fig. 8 1, a larva corresponding in age approximately 
to that of Fig. 74 is represented as seen from the left side. 

As noted above, the posterior primary slits bend nor- 
mally under the pharynx at this stage, and some of them 
extend as much on one side of the body as on the other, 
being continued across the ventral side of the pharynx. 
The great feature of this larva is the relatively prodigious 
mouth, through which the upper portions of the first four 
primary slits can be seen. 

From this side we look into the depths of the praeoral 
pit, having only seen it by transparency in the preceding 

Fig. 81. Anterior portion of larva, with thirteen gill-slits, from the left side. 
(After WILLEY.) 

olf. Olfactory pit, communicating with neuropore. x. " Nephridium " of Hat- 
schek. n.t. Spinal cord. ch. Notochord. p.p. Prseoral pit. ex. External open- 
ing of club-shaped gland, ci. Rudiment of buccal cirri, p.b. Peripharyngeal band. 
m. Mouth, /.ji 3 . Thirteenth primary slit. 

figures. It is continued backwards into a ciliated groove, 
which abuts on the dorsal margin of the mouth. Prob- 
ably most of the food which enters the mouth passes 
along this groove. 

Below the pointed anterior extremity of the mouth is to 
be seen the external aperture of the club-shaped gland, 
and a short distance behind this is a round, refringent 
body, which has become differentiated from the gelatinous 


connective tissue lying below the epidermis, and repre- 
sents the rudiment of the first element of the cartilagi- 
nous skeleton of the buccal cirri. 

Running parallel with the lower margin of the mouth, 
and curving gently upwards to the dorsal wall of the 
pharynx, is a ciliated band proceeding from the lower limb 
of the endostyle, and corresponding to the one on the other 
side, which we found in connexion with the upper portion 
of the endostyle. Its course on the left side is somewhat 
different anteriorly from that of the right side, owing to 
the position and size of the mouth. (Cf. Figs. 78 and 81.) 

The so-called olfactory pit, which arose at a much earlier 
stage as an ectodermic depression above and in connexion 
with the neuropore, no longer lies in the mid-dorsal line as 
in Fig. 64, but it has been displaced to the left side by the 
upgrowth of the dorsal fin (Fig. 81). Here, as in the case 
of the anus, the development of a median fin has no other 
effect on the aperture in question than to cause it to 
forsake its primitively median and symmetrical position 
and to assume an asymmetrical position on the left side of 
the body. This is important to bear in mind, as the asym- 
metrical position of the mouth will be explained below on 
an analogous basis. 

For the present it is sufficient to call attention to the 
fact that, with the exception of the gill-slits, whose primary 
unpaired character is due to the retarded or latent develop- 
ment of their antimeres, the unpaired openings in the 
body-wall namely, neuropore, praeoral pit, external aper- 
ture of club-shaped gland, mouth, and anus all lie on the 
left side of the body. 

At a slightly later stage than the preceding, the front 
end of the mouth is found to be no longer pointed, but to 
have become rounded off, and, moreover, to lie at a deeper 


level than previously (Fig. 82). The posterior groove of 
the prasoral pit which we described in the last stage, seems 
to be preparing the way for the mouth to dip inwards 
towards the right wall of the pharynx, which, in fact, it has 
actually begun to do. 

At a still later stage, corresponding to that shown in 
Fig- 77> tne shape of the mouth has become entirely altered 
(Fig. 83). 

It has now the form of a triangle with the apex directed 
backwards and the base standing vertically in front. But 
the apex and the base are not in the same tangential plane, 

Fig. 82. Anterior portion of larva somewhat older than preceding, to show 
commencing adjustment of the mouth. (After WlLLEY.) 

e. Endostyle seen through the mouth. Other letters as above. 

the former being on the left side of the body, and the latter 
much deeper inwards ; in fact, just below the skin on the 
right side of the body. (Cf. Fig. 77.) 

We see, therefore, that the longitudinal diameter of the 
larval mouth is gradually shortening. It is eventually 
reduced to zero when the right and left sides of the mouth 
or velum come to lie opposite to one another, the velum 
ultimately attaining a circular form and a median sub- 
vertical position underneath the oral hood. When the 
larva has reached the age to which Fig. 1 1 refers, the right 


half of the velum is nearly but not even yet quite opposite 
to the left half (Fig. 93). 

In the preceding stage (Fig. 82) there were several 
additional buccal cartilages added to the first one which 
we described. In the present stage these have begun to 
grow outwards so as to produce small notches in the 
integument, which is now commencing at this point to 
form the right half of the oral hood. The left half of the 
latter arises as a downgrowth of the integument from the 
upper margin of the prseoral pit and its posterior continua- 
tion, the above-mentioned ciliated groove. (Cf. Figs. 81, 
82, and 83.) The hinder portion of this fold is at first on 


Fig. 83. Anterior portion of still older larva, from the left side, to show 
change in shape and position of the mouth. (After WlLLEY.) 

Letters as above. The left half of the oral hood is now growing down over the 
mouth and prasoral pit. 

a level with the dorsal margin of the mouth, and in fact 
merges into the latter, but subsequently grows over it, 
extending to its posterior extremity, where it meets the 
right half of the oral hood. 

It is obvious from the above description and figures that 
a large part of the right wall of the oral hood is derived 
from the original wall of the snout below the praeoral pit, 
and so an explanation is afforded of the fact noted in the 
first chapter that the right half of the oral hood is continu- 
ous round the anterior extremity of the notochord with 
the cephalic expansion of the dorsal fin. 9 



The prseoral pit itself is absorbed, as it were, into the 
oral hood, so that it eventually loses its independent exist- 
ence as a pit, although the sense-organ of the praeoral pit 
persists in the adult as a deep groove in the dorsal wall 
of the oral hood to the right of the base of the notochord. 
The remaining ciliated epithelium of the original praeoral 
pit increases in extent, and grows out into the finger- 
shaped tracts which we have already described as being 
characteristic of the inner surface of the oral hood, consti- 
tuting the so-called " Raderogan." (Cf. Fig. 3.) 

Equalisation of the Gill-slits. 

In the stage next succeeding that of which a ventral 
view is given in Fig. 80, the first eight primary slits that 
is to say, from the original second to the ninth inclusive - 

Fig. 84. Larva toward the close of the metamorphosis, from the left side. 
(After WILI.EV.) 

o. Olfactory pit. v. Velum, p.b. Peripharyngeal band. e. Endostyle. /.A Second 
primary slit, the first having closed up. m. Left metapleur. s.a. Floor of atrium. 
p,s vl . p.s 13 . Vestiges of the twelfth and thirteenth primary slits. 

have become definitely established on the left side, their 
longitudinal and vertical axes are equalised, and in most 
of them the tongue-bars are completely formed (Fig. 84). 
No tongue-bar is formed in the first slit on either side, and 
this slit apparently remains as a rule simple throughout 


In Fig. 84 the last indications of the twelfth and thir- 
teenth primary slits are to be observed as slight depres- 
sions in the floor of the pharynx in the mid-ventral line. 
The tenth and eleventh slits would close up later. 

It should be pointed out that the closure of the poste- 
rior primary slits does not proceed in perfect correspond- 
ence with the age of the larva, but takes place sometimes 
at an earlier and sometimes at a later stage than here 

The gill-slits of both sides now begin to elongate in 
the vertical direction (Fig. 93), and eventually a very well- 
marked stage is reached, which is characterised by the 
presence of eight pairs of gill-clefts. This latter stage 
would appear to have a considerable duration, and, as it 
stands on the borderland between the larva and the adult, 
and forms the commencement of what may be called the 
adolescent period of the development, it may well be 
regarded as a critical stage. By this time the young 
Amphioxus has given up its free pelagic life in the open 
sea, and has commenced to burrow in the sand, which it 
continues to do for the rest of its life.* 

Further Growth of Endostyle, etc. 

At the point at which we left the endostyle, its two 
halves were in the relation to one another of upper and 
lower. During the steps in the metamorphosis which we 
have recorded above, the upper half of the endostyle is 
brought down to the same level as the lower half on the 
right side of it, and so the definite form of the endostyle 
is established by the conjunction of its right and left 
halves. It then proceeds to grow backwards along the 

" The duration of the larval development of Amphioxus may be estimated 
at about three months. 


base of the pharynx between the two rows of gilt-slits, 
but does not reach the posterior end of the pharynx until 
a much later period. 10 

The features in the development of the endostyle which 
ought to be especially emphasised are, firstly, its direc- 
tion of growth from before backwards, and secondly, its 
primary anterior position in the wall of the pharynx in 
front of all the gill-slits. 

In connexion with the modification in the shape and 
position of the mouth, as described above, it is important 
to insist on the fact that the mouth of the larva is directly 
converted into the velum of the adult, while the oral hood 
which grows over the mouth is a new formation. 

During the period of the metamorphosis the larva does 
not increase in length. It is rather a readjustment of 
parts which is then taking place than an increase in bulk 
which is the symbol of active growth. From the time of 
the first indication of the secondary slits (Fig. 74) till 
after the completion of the passage of the primary slits 
from the right to the left side of the body, the average 
length of the larva may be taken as approximately 
3.5 mm. 

The adolescent period is essentially the period of active 
growth in bulk and maturity. The increase in length 
during this period does not, however, depend on the 
addition of new myotomes to those already formed, but 
merely on the progressive growth in size of the latter. 
The full complement of myotomes was developed during 
the early larval period, and is present in the larva repre- 
sented in Fig. 74. 


Development of Reproductive Organs. 

One of the most interesting events which we have now 
to chronicle is the development of the reproductive organs. 
This commences when the young Amphioxus has reached 
the length of about 5 mm. 

Our knowledge of the details of the processes involved 
in the formation of the genital organs is again due to the 
work of BOVERI, who has made the discovery that the 



Fig. 85. Transverse section through the pharyngeal region of a young 
individual of 5 mm., to show place of origin of sexual elements. (After BOVERI.) 

f. Fascia, e.c. Portion of ccelom, which will form the endostylar coelom. 
tig. Primitive sexual cells in the lower angle of the myoccel. Other letters as above. 

primitive sexual cells arise in the cavity of the myotome 
by differentiation of certain of the epithelial cells lining 
the myocosl. 

It had previously been assumed that they were derivatives 



of the peritoneal epithelium lining the general body-cavity. 
The fact that they arise in the way shown by Boveri is one 
of great morphological importance. 

In a transverse section of a young individual 5 mm. 
in length, the primitive sexual cells are to be recognised 
as a closely packed group of cells, with large nuclei in the 
lower angle of the myotome ; that is, in the angle formed 
by the membrane which divides the myoccel from the 
splanchnocoel, which we may call the intcrccelic membrane, 
with the cutis (Fig. 85). Since the myotomes of one side 
alternate with those of the other, so do the centres of 

Fig. 86. Longitudinal views of the developing gonads, obtained by dissecting 
out the ventral borders of the myotomes. (After BOVERI.) 

u.g. Primitive sexual cells arising from the myocrelic epithelium ; the nuclei 
scattered about the surface of the preparations also belong to the myocoslic 

formation of the primitive sexual cells, and in a given 
section, as in Fig. 85, only one such centre is to be observed 
on the right or left side of the section, as the case may be. 
Its actual position in the longitudinal aspect of the myo- 
tome is shown in Fig. 86 A, B, and C. The formative 
centres of the primitive sexual cells lie at first in the angle 
mentioned above, but applied to the posterior faces of the 
dissepiments between the myotomes (Fig. 86 A). 

At a somewhat later stage, having slightly increased in 
bulk, they begin to push the dissepiments before them 



so as to make a projection into the myocoel in front (Fig. 
86 B, C}. This projection of the primitive gonad into the 
myocoel next in front of that to which it originally belonged, 
is gradually carried to such an 
extent that the gonad becomes 
entirely shut off from its original 
myocoel and hangs freely into the 
next one, being connected by a 

Short Stalk with the anterior face Fig. 87. Similar prepara- 
tion as the preceding, showing 

of the dissepiment and surrounded a later stag e i n the development 
by a membrane which is obviously B ^ e R j mitive gonad ' (After 
derived from, and for some time 

continuous with, the original dissepiment (Fig. 87). In 
correlation with the increase in size of the primitive gonad, 
an evagination of the basal wall of the myocoel in which it 
now lies, takes place, and by the time the young Amphi- 

Fig. 88. Preparation showing the rhomboidal pouches of the myoccel 
which project into the atrial cavity. (After BOVERI.) 
This condition is found in individuals of 13-14 mm. 

oxus has attained a length of 13 or 14 mm. there is, in 
connexion with each primitive gonad, a wide rhomboidal 
expansion of the lower portion of each corresponding 
myoccel projecting into the atrial cavity (Fig. 88). 

The cavity of these sacs, to the wall of which the gonads 
are at this stage still united by a stalk, constitutes the so- 
called perigonadial cceloni^- or cavity of the gonadic 
pouches, which, at the time of sexual maturity, is entirely 
filled up by the sexual elements. 



The gonadic pouches next become gradually constricted 
off from the myoccelic spaces, and eventually lose all com- 
munication with them. In the midst of the at first solid 

mass of primitive sexual 
cells a cavity subsequently 
appears, and the gonad be- 
comes a hollow sac (Fig. 89). 
In the course of its fur- 
ther growth the gonadic sac 
(not to be confused with the 
gonadic pouch in which it 
lies) grows out into a num- 

Fig. 89. Portion of transverse sec- i f i i i 

tion through an individual of 13 mm., Der lappets, and SO D6- 

to explain the conditions observed in CO meS a racemose reproduc- 
preceding preparation. (After BOVERI.) 

b.v. Blood-vessel, go. Gonadic sac. tive gland (Langerhans). 

p.g.c. Perigonadial ccelom (sonadic .-pi ... , ,, 

pouch). ^.Transverse muscles. The The primitive SCXUal cells 

index line to which there is no letter remain for a considerable 

indicates the fold by which the gonadic 

pouch becomes constricted off from the length of time in an absO- 

lutely indifferent condition, 

and it is impossible to distinguish the male from the 

According to LANGERHANS, sexual differentiation does 
not begin to take place until the individuals have reached 
a length of 17 mm., and sometimes it does not occur until 
a much later period. It is inaugurated by the commence- 
ment of the processes of spermatogenesis and ovogenesis. 
There are no accessory sexual characters in Amphioxus, 
and the sex can only be determined by an examination of 
the reproductive glands. 

The segmental arrangement of the formative centres 
of the reproductive organs at the base of the myotomes 
is again met with in the embryonic development of the 
Selachians, as shown by RUCKERT (Fig. 90). Here, also, 



the primitive sexual cells 
make their first appearance 
in the segmented area of 
the trunk at the base of 
the somites. Later on, by 
differential growth, they 
come to lie on the dorsal 
wall of the unsegmented 
peritoneal cavity, and their 
primitive segmental origin 
is entirely obscured ; while 
in Amphioxus the primitive 
segmentation of the gonads 
is maintained throughout 

This forms another most 
interesting example of the 
way in which the adult 
Amphioxus, in the details 
of its organisation, essen- 
tially resembles the em- 
bryos of the higher types. 

Fig. 90. Horizontal section through 
the ventral portion of six consecutive 
mesodermic somites of an embryo of 
Pristiurus, to show the segmental origin 
of the sexual elements. (After RUCKERT). 

c. Cavities of somites, g.c. Sexual 

This observation of Riickert's has 
recently been doubted, with how much 
justice it is difficult to say, by MlNOT 
(Gegen das Gonotom. Anat. Anz. IX. 
1894. pp. 210-213). 


We will now pass on to give a general interpretation of 
some of the principal phenomena which are presented to 
us in the development of Amphioxus. 

Larral Asymmetry. 

By far the most prominent feature of the fully formed 
larva is its astounding asymmetry, and it is extremely 
important, from a morphological point of view, to form a 
just conception of it. 


The phenomenon of asymmetry manifests itself in the 
larva of Amphioxus under several very different aspects, 
and is occasioned by various causes. For convenience we 
may classify the forms of asymmetry which we have to 
consider under three main divisions, according to the type 
of organs involved. 

1. Median Asymmetry. This relates to such structures 
as lie normally in the middle line, whether dorsal or ven- 
tral, but which have been mechanically or correlatively dis- 
placed from their primitive position by the differential 
growth of neighbouring parts. Such are the olfactory pit 
and neuropore, the anus, the mouth, and the endostyle. All 
these are essentially and primordially median and unpaired 
structures. We have already dealt with the neuropore 
and anus, while the mouth and endostyle will be con- 
sidered below. 

2. Bilateral Asymmetry. This refers to the alternation 
of paired structures, such as myotomes, spinal nerves, gill- 
slits, and gonads, which we have already noted in the adult 
organisation. Primarily the organ of one side lies opposite 
to its antimere of the other side. By a secondary displace- 
ment it comes to alternate with it.* 

3. Unilateral Asymmetry, --Next to the asymmetrical 
mouth, this is perhaps the most striking form of asym- 
metry which the larva of Amphioxus exhibits. It relates 
to those structures which belong to the category of paired 
organs, but which, in the course of the larval development, 
appear unpaired on one side of the body. Such are the 

* When the myocoelomic pouches first appear in the embryo they are 
placed symmetrically. At an early stage, however (see Fig. 63 .5), the alter- 
nation sets in. This involves such later-appearing structures as the spinal 
nerves and gonads, so that they alternate from the time of their first origin. 
The alternation of the gill-slits would seem to be independent of that of the 


gill-slits and the praeoral pit. As described in the fore- 
going pages the asymmetry of the praeoral pit is a second- 
ary occurrence, since it arises at first as one of a pair of 
symmetrically disposed head-cavities, or anterior intestinal 
diverticula, while the unilateral asymmetry of the gill-slits 
is ontogenetically primary. The unilateral gonads of the 
species of Amphioxus from the Bahamas and Torres 
Straits also belong to this category. 

Although, on account of their essentially azygous nature, 
the mouth and endostyle have been separated from the 
gill-slits in the above classification, it is obvious that their 
asymmetrical position in the larva must be ascribed to one 
and the same cause. In the succeeding pages we shall 
endeavour to demonstrate what this cause was. 

Explanation of Asymmetry of Mouth and Gill-slits. 

It is quite evident that the primary gill-slits which 
appear on the right side of the larva belong primitively, 
or ancestrally, to the left side, to which, in fact, they are 
eventually transferred. Meanwhile, the left side of the 
larval pharyngeal region is largely occupied by the huge 
oral aperture. 

We may figure to ourselves the primitively left-side gill- 
slits being carried over to the right side by a semi-rotation 
from left to right of the pharyngeal wall. The primitive 
right side of the pharynx would thus be crowded out, so to 
speak, and the right-side gill-slits would be temporarily 
obliterated owing to lack of room, while the original mid- 
ventral line would be carried high up on the right side, 
where, in point of fact, it is plainly indicated by the bran- 
chial artery, which lies actually above the primary gill-slits 
in the larva (Fig. 74, etc.). 


Thus the actual topographical conditions in the larva do 
not by any means coincide with the morphological rela- 
tions of parts, since the morphological mid-ventral line of 
the pharynx lies high up on the right side of the body. It 
should be carefully noted that the form of asymmetry 
which we are now considering only affects the anterior 
portion of the larval body. 

The same semi-rotation of the pharyngeal region which 
converted the primitive left side of the larva into the 
actual right side caused the primitively median mouth to 
take up its position on the actual left side. But since, as 
we have noted, the rotation occurred from left to right, 
the mouth must have been originally situated in the 
median dorsal line. 

In postulating a virtual semi-rotation of the ancestral 
pharynx, we do not, of course, mean to suggest the prob- 
ability of an actual movement in bulk about the longi- 
tudinal axis, but merely that the formative centres of the 
various structures belonging to this region of the body 
(gill-slits, mouth, endostyle, etc.) have, by the correlated 
interaction of their component cell-groups, been diverted 
from their ancestral relations through the intercalation, in 
the course of the progressive evolution of the organism, of 
a new and disturbing element. 

We are now in a position to say what this disturbing 
element is. It is the secondary forward extension of the 
notochord beyond the limits of the dorsal nerve-tube to 
the tip of the snout. As already stated, there is direct 
evidence to show that this is a secondary and not an an- 
cestral feature, inasmuch as in the young embryo (Fig. 
68 bis} the notochord is removed from the anterior extrem- 
ity of the body by a very appreciable interval, which is oc- 
cupied by that portion of the archenteron which gives rise 


to the head-cavities. Moreover, as was pointed out above, 
the dorsal groove of the archenteron, which gives rise to 
the notochord, remains open into the archenteric cavity 
in the region of the first myotome, and even somewhat 
behind the level of the neuropore, for some time after 
its walls have approximated to form the solid notochord 
behind this region. 

The forward extension of the notochord in Amphioxus 
is, therefore, dc facto, to a large extent an ontogenetic 
phenomenon, although, from the very beginning, it shows 
what may be described as a precocious tendency to extend 
beyond the nerve-tube. We shall also find that there is 
every reason to suppose that it is a cenogenetic, and not a 
palingenetic, feature. 12 

Since we know for an actual fact that the primary gill- 
slits of the larva belong ancestrally to the left side, it fol- 
lows as an absolute topographical necessity that the mouth 
has been brought to one side from an originally median 
dorsal position, by the same semi-rotation of the pharynx 
(in the sense explained above) which has demonstrably 
carried the primitive left-side gill-slits under the pharynx 
up to the right side of the larva. But this is not the only 
criterion by which we can judge of the ancestral position 
of the mouth. 

In the larvae of the Ascidians, the nearest existing rel- 
atives of Amphioxus, there is a proeoral lobe and a neuro- 
pore, which opens at first to the exterior in the mid-dorsal 
line, just as in Amphioxus. But in contrast to the latter 
form the notochord does not extend forwards into the re- 
gion of the praeoral lobe, but it stops short behind the 
cerebral vesicle. 

Immediately in front of the neuropore, in the Ascidian 
larva, the wall of the pharynx comes into contact with the 


ectoderm and fuses with it, and then at the point of fusion 
a perforation takes place, and the mouth is established in 
the mid-dorsal line. During the formation of the mouth 
the neuropore temporarily closes up, but subsequently it 
reopens- -into the moutli, 

In Amphioxus we can only assume that in correlation 
with the forward extension of the notochord, the mouth 
was compelled to forsake its primitive relations to the 
neuropore and to move to one side so as to make way for 
the notochord. The growth of the latter to the front end 
of the body obviously prevents the wall of the pharynx 
from coming into contact with the ectoderm in the mid- 
dorsal line, while it leaves the neuropore unaffected, since 
the nerve-tube is essentially dorsal to the notochord, and 
the pharynx, on the other hand, essentially ventral to it. 

This explains the fact that the hypophysis (olfactory pit) 
of Amphioxus opens dorsally directly to the exterior instead 
of into the mouth as it does in the Ascidian. 

The secondary gill-slits that is, those belonging to the 
primitive right side of the body present an interesting 
instance of retarded or latent development. This is due 
to the fact that their own side of the body is at first 
usurped by their primitive antimeres, the so-called primary 
slits, as a result of which they have themselves been 
temporarily crowded out as mentioned above. In con- 
sequence of their retardation, when they do appear to 
inaugurate the process of symmetrisation, they do not 
conform to the method in which metameric structures are 
normally produced, but most of them namely, from the 
second to the seventh inclusive arise simultaneously 
while the first and the eighth arise somewhat later. 


Larval Asymmetry not Adaptive and not Advantageous; 
Forward Extension of Notochord Adaptive and Advan- 

The conclusion to be drawn from the above considera- 
tions is that the remarkable asymmetry of the larva of 
Amphioxus, in respect of the pharynx and the parts con- 
nected with it, is of no specific advantage whatever to the 
larva, but is merely a stage, which has been preserved in 
the ontogeny, of a topographical readjustment of parts 
necessitated by the removal of the mouth from its primi- 
tive mid-dorsal position in consequence of the secondary 
forward extension of the notochord, which has thus caused 
a virtual semi-rotation of the pharyngeal region of the 
body. On the other hand, the forward extension of the 
notochord is a distinct advantage in later life, since, by 
giving resistancy to the snout, it enables the animal to 
burrow its way into the sand with such astonishing facility, 
while the fact that it grows to the front end of the body at 
a very early stage in the embryonic development, long 
before it comes to be put to this definite use, must be 
regarded as an instance of precocious development of which 
there are numerous and otherwise inexplicable examples 
in the field of comparative embryology. 

The larval asymmetry of Amphioxus is therefore a purely 
secondary or cenogenetic feature, and has no directly ances- 
tral or palingenetic significance, although, as we have shown 
above, it serves indirectly as a clue to what the ancestral 
condition was. At the same time it is a primary feature 
in the actual ontogeny ; that is to say, the asymmetrical 
structures (mouth and gill-slits) arise in situ, and are not 
removed in the individual development from a primary 


symmetrical to a secondary asymmetrical position, as is the 
case, for instance, with the neuropore. 

It may appear paradoxical, but is nevertheless correct, to 
say that in the ontogeny the mouth and gill-slits appear 
Primarily in a secondary position. 

It is quite evident that the asymmetry of the larva of 
Amphioxus is of a totally different character to the well- 
known asymmetry of the flat-fishes or Pleuronectida 
(turbot, sole, plaice, halibut, flounder, etc.). The latter 
are hatched as perfectly symmetrical larvae with eyes quite 
opposite to one another. Then, in adaptation to a life at 
the bottom of the sea, after a short pelagic existence they 
turn over on one side, in some species the right side, and 
in others the left, and the eye of that side moves over the 
snout, sometimes even through the snout, to the other 
side, and so the eyes come to lie on the same side. In this 
case, therefore, the asymmetry, which is secondary in every 
sense of the word, is the result of a special adaptation to a 
particular habit of life, and is accordingly of the greatest 
advantage to the fishes which possess it. 

On the other hand, its extraordinary asymmetry is of 
no conceivable advantage to the larva of Amphioxus, and 
does not represent an adaptation to any peculiar mode of 
existence whatever. 

It is rather the mechanical, incidental, accessory, and 
subsidiary accompaniment of another organic change which 
is both advantageous and adaptive, namely, the forward 
extension of the notochord ; and while the excessive asym- 
metry is indifferent to the pelagic larva, it would be posi- 
tively detrimental to the adult. 

Thus in all respects the larval asymmetry of Amphioxus 
is the precise converse of the adult asymmetry of the 
Pleuronectidae. 13 



We will now pass on to consider what new light the 
larval development of Amphioxus throws on its relation- 
ship to the craniate Vertebrates. 

As a type of the latter with which to make the com- 
parison, we will select Ammoccetes, the larva of the lamprey, 
Petromyzon, which is the nearest relative of Amphioxus 
among the Craniota. 

Nervns BrancJiialis Vagi. 

Although Ammoccetes possesses an organisation which, 
especially in virtue of its nervous system and sense- 
organs, entitles it to an undoubted place among the 
craniate Vertebrates, yet, on the whole, its structural ele- 
ments remain in such a relatively simple condition of 
elaboration that it readily adapts itself to a comparison 
with Amphioxus. 

At the same time the system of ganglia and peripheral 
cranial nerves indicated in Fig. 91 will show what a great 
gap there is between the two forms. Nevertheless, a 
nerve corresponding to that which lies over the gill-slits 
in Fig. 91, the ncrvus brancJiialis vagi, has recently been 
discovered in Amphioxus by VAN WIJHE, so that there 
need be no difficulty in comparing the pharyngeal tract 
of Ammoccetes with that of Amphioxus. 

It may be added here that the nerve-supply of the 
pharynx of Amphioxus was described as a branchial plexus 
by ROHON and FUSARI, but the origin of the nerves which 
gave rise to the plexus was not satisfactorily determined, 
beyond the fact that they arose from the rami viscerales 
of the dorsal spinal nerves. VAN WIJHE also was not 



able to determine the precise origin of the longitudinal 
nerve discovered by him. This nerve, which lies on either 
side at the place where the ligamentum denticulatum 
passes into the gelatinous lamella derived from the inter- 
ccelic membrane, gives off the branches which form the 
"branchial plexus." Van Wijhe states that the origin of 
the " ramus branchialis vagi " of Amphioxus is to be 
sought in the eighth to the tenth dorsal spinal nerves. 

Fig. 91. Anterior portion of young Ammoccetes of 4 mm., to show extension 
of brain, origin of endostyle (thyroid), relations of branchial nerves, etc. (After 


/, 11, III, IV. The so-called " Hauptganglia." / and //. Trigeminus. 
///. Acustico-facialis. IV. Glossopharyngeus. V. Vagus. 

au. Auditory capsule, ch. Notochord. e. Endostyle (hypobranchial groove, 
thyroid), hy. Hypophysis, in front of which is the nasal groove, n.l. Nervus 
lateralis. Nervus branchialis. o.p. Eye. /. Pineal body (epiphysis). 
p.m. Prasoral endodermic pouch (median portion of praemandibular cavity. 
st. Stomodceum. /, VIII. First and eighth gill-pouches; the small circles behind 
the gill-pouches indicate the positions of the external openings of the gill-pouches, 
which will become perforated later. The small black spots in front of the (later 
appearing) external openings represent the so-called ganglia prcztrematica. 

He found that the nerve curved ventralwards in front and 
passed downwards through the interccelic membrane until 
it reached the level of the ventral transverse muscles in 
front of the visceral branch of the eleventh spinal nerve. 
He was unable to follow it further in the complex nerve- 
plexus which lies on the surface of the muscles. It is 
probable, however, that the branchial nerve arises from 


the visceral branch of the eighth, ninth, or tenth spinal 

Stomodceum, Hypophysis, and Gill-slits. 

It is a common fact that the time and order of forma- 
tion of corresponding parts differ greatly in the develop- 
ment of different species. Thus in Ammoccetes, at the 
stage shown in Fig. 91, the definitive mouth, correspond- 
ing to the velum in Amphioxus, has not yet formed, but 
the equivalent of the oral hood is already present in the 
form of a deep in-pushing of the ectoderm which, at its 
blind end, is closely applied to the anterior endodermic 
wall. The mouth will break through later in the middle 
of the area of contact between ectoderm and endoderm. 

This ectodermic invagination, whose cavity is probably 
the homologue of the vestibule formed by the oral hood 
which leads into the mouth in Amphioxus, is known as 
the stomodaeum. Immediately in front of the stomodceum 
is another ectodermic involution which is in contact with 
the front of the brain, and is known as the hypophysis or 
pituitary body^ 

It will appear later that this is the probable equivalent 
of the so-called olfactory pit of Amphioxus. 

In the wall of the pharynx of Ammoccetes there are, at 
this stage, the indications of eight pairs of gill-slits. They 
have not yet, however, broken through to the exterior, but 
consist of a succession of hollow outgrowths of the phar- 
ynx stretching towards the ectoderm with which they will 
eventually fuse (Fig. 92 A, B, C). 

In the case, however, of the first pair of gill-pouches, 

* It is not impossible that many of the rami viscerales may send up branches 
to the branchial plexus, as was indeed described by Rohon. In this case, 
Van Wijhe's nerve would be of the nature of a collector. 

1 66 


it does not come to a fusion with the ectoderm ; but in- 
stead they begin to undergo a retrogressive development 
and eventually flatten completely out (Fig. 92 B}. They 
are thus shown to be rudimentary structures, morphologi- 
cally representing the first pair of gill-clefts, but never 
achieving their full development. 

Fig. 92. Horizontal sections through the pharyngeal region of Ammoccetes, 
to show the relation of the first pair of gill-pouches to the peripharyngeal grooves. 
(After DOHRN.) 

A. Two days after hatching ; first pair of gill-pouches well developed. 

B. Six days after hatching ; first pair of gill-pouches flattened out. 

C. Nine days after hatching ; appearance of peripharyngeal grooves. 
I-VIIl. Gill-pouches, b.w. Body-wall, oes. Oesophagus, ph. Pharynx. 

ph.g. Peripharyngeal groove, st. Stomodoeum. vel. Velum. 

As to their position, they occupy the extreme anterior 
angles of the pharynx formed by its lateral walls with the 
anterior transverse wall against which the Stomodoeum is 
applied. Whatever may be the reason for it, the atrophy 
of the first pair of gill-pouches in Ammocoetes is of pre- 
cisely the same nature as the atrophy of the first gill-slit 
in Amphioxus, with the distinction that the latter actually 
opens to the exterior for a time. 


Endostyle or Hypobrancliial Groove. 

At a stage in the development of Ammocoetes which 
precedes the flattening out of the anterior gill-pouches, 
a median depression occurs in the extreme anterior 
region of the ventral wall of the pharynx between the 
first pair of gill-pouches. In its production the wall of 
the pharynx at this region projects itself ventrally and 
slightly forward. This groove, which is known as the 
hypobmnchial groove, develops in the direction from before 
backwards, and eventually extends backwards as a longi- 
tudinal groove as far as the fifth pair of gill-pouches 
(Fig. 91). 

WILHELM MULLER showed that it was the homologue 
of the endostyle of Ascidians and Amphioxus, and he has 
been amply confirmed by DOHRN. It agrees with the lat- 
ter structure in its origin at the anterior extremity of the 
pharynx and subsequent growth backwards and in its 
histological structure, the most marked feature of the lat- 
ter being the four longitudinal rows of gland-cells which 
were noted above in the endostyle of Amphioxus. (Cf. Fig. 
13.) Like the latter, also, it is a slime-secreting gland. 

In Ammocoetes the hypobranchial groove becomes 
largely shut off from the cavity of the pharynx by the 
gradual ingrowth of a diaphragm-like lamella which pro- 
ceeds from the angle made by the groove in front with the 
anterior wall of the pharynx (Fig. 91). Subsequently a 
similar diaphragm grows in from the posterior margin of 
the groove, and finally the latter only communicates with 
the pharynx by a small aperture in the mid-ventral line 
between the fourth pair of gill-pouches. 

1 68 


P eripharyngeal Ciliated Bands of Ammocoetes. 

Corresponding with the right and left peripharyngeal 
ciliated bands which we described as proceeding from the 
anterior borders of the endostyle in Amphioxus there is 
a pair of ciliated grooves in the pharyngeal wall of Ammo- 
ccetes which proceed from the anterior lip of the hypo- 
branchial groove after the latter has become to a large 
extent shut off from the pharynx by the above-mentioned 
diaphragm. These grooves curve forwards and upwards 

Fig. 93. Young Amphioxus, after the metamorphosis, having eight gill-slits 
on each side. From the right side. (After WlLLEY.) 

p,b. Peripharyngeal band. v. Velum ; shown separately below the main figure, 
with rudiments of four velar tentacles, e. Endostyle, extending backwards to the 
level of the fourth gill-slit, r.m. Right metapleur. 

in front of the gill-clefts (after the obliteration of the first 
pair of gill-pouches), and then proceed backwards on either 
side of the dorsal middle line of the pharynx as far as the 
commencement of the oesophagus. Here they appear to 
curve downwards again, and uniting together, extend for- 
wards as a median ventral groove to the posterior lip of 
the hypobranchial aperture. 


The last-mentioned median ciliated groove would appear 
to be unrepresented in Amphioxus, but the downward 
curvature of the ciliated bands of the latter behind the gill- 
slits can be observed (Fig. 93). 

In Ammoccetes the ciliated peripharyngeal grooves, 
where they curve upwards in front along the anterior wall 
of the pharynx, apparently occupy the same position which 
was previously occupied by the first pair of gill-pouches. 
Since the latter have already entirely disappeared, there is 
nothing in the way of their occupying this position (Fig. 
92 C). In Amphioxus, where the corresponding gill-slit 
remains open for a long time, the peripharyngeal band exists 
without connexion of any sort with the portion of the wall 
occupied by the slit, and when the latter closes up, it leaves 
no trace behind. 15 

Thyroid Gland. 

When the metamorphosis of Ammocoetes into Petromy- 
zon takes place (which happens after a larval existence of 
some two years' duration), the hypobranchial groove loses 
all connexion with the pharynx and becomes broken up 
by the ingrowth of connective tissue into a number of 
separate capsules which collectively constitute the thyroid 
gland of Petromyzon. 

The thyroid gland is one of those enigmatical ductless 
glands which form such a curious and constant feature of 
the Vertebrate organisation. 

There is considerable doubt as to the specific physio- 
logical function which it has to perform, but at the same 
time it is a necessary factor in the Vertebrate economy, 
and is of great importance from a pathological point of view. 

In the higher forms it is attached to the lower side of 
the larynx, and appears to have received its name on 


account of its close proximity to the thyroid cartilage of 
the latter, the older anatomists assuming a functional 
relation between the two structures. 

We know perhaps more about the morphological than 
about the physiological significance of the thyroid gland, 
since it is the vestige of the very actively functional endo- 
style or hypobranchial groove of the Ascidians, Amphioxus, 
and Ammocoetes. 

Morphology of Club-shaped Gland of Amphioxus. 

In describing above the formation of the second row of 
gill-slits in Amphioxus, we found that the first secondary 
slit paired with the second primary slit. It now remains 
to consider what has become of the antimere of the first 
primary slit. 

The probability is that, unlike the antimeres of the suc- 
ceeding primary slits, that of the first has not suffered a 
retardation of development, but is present from the very 
beginning of the larval development, although in a some- 
what modified form. I refer to the club-shaped gland. 

The club-shaped gland fulfils the requirements of a gill- 
slit in so far as it opens at one end into the pharynx, and 
at the other to the exterior. Since, as we have shown, the 
morphological mid-ventral line lies high up on the right 
side, immediately above the primary gill-slits, it is evident 
that its anterior continuation would pass through the en- 
dostyle precisely at the point where the latter is redoubled 
upon itself. But the internal opening of the club-shaped 
gland lies above the upper limb of the endostyle, and 
therefore it is placed not only on the actual right side of 
the larva, but in opposition to the first primary slit, on the 
morphological right side as well. 


It must be supposed that the original gill-slit, from 
which the club-shaped gland is derived, acquired, for some 
reason or other, a tubular form. 

A familiar precedent for gill-slits being drawn out into 
elongated tubes, the effect of which is to separate the in- 
ternal from the external opening by a long interval, is 
presented by the hag-fish, Myxine. Myxine also shows us 
that, in correlation with the canalisation of the gill-slits, 
their external apertures may enter into new relations dif- 
fering considerably from the primitive condition. As is 
well known, the elongated tubular gill-clefts of Myxine do 
not open separately to the exterior, but fuse together at 
their distal extremities, so as to give rise to a longitudinal 
duct on each side, which opens to the exterior some dis- 
tance behind the gill-region. 

It is only on some such supposition as this namely, that 
the external aperture of the gill-slit represented by the 
club-shaped gland of Amphioxus has assumed new topo- 
graphical relations in correlation with the canalisation of 
the original slit that its position on the opposite (left) side 
of the body to the internal opening of the gland is ren- 
dered intelligible. The position of the internal opening 
furnishes the criterion by which to judge of the primitive 
relations of the original gill-slit. 

With the above point of view, therefore, we may signal- 
ise the following facts to prove that the club-shaped gland 
is the antimere of the first primary gill-slit. 

1. They arise simultaneously in the embryo as grooves 

in the ventral wall of the pharynx. 

2. They come to lie on opposite sides of the morphological 

median line the first gill-slit entirely so, and the 
club-shaped gland in respect of its internal opening 
into the pharynx. 


3. They atrophy and disappear simultaneously during the 

metamorphosis of the larva. 

4. No secondary gill-slit ever arises to pair with the first 

primary slit. 

As the stage represented in Fig. 64 marks such a vital 
turning-point in the development of the individual, being 
the stage at which the embryo becomes a larva and the 
struggle for existence in obtaining independent nourish- 
ment genuinely sets in, it is important to be able to define 
it accurately. In view of the above considerations, we 
arrive at the conclusion that the larva is at this stage 
possessed morphologically of a pair of gill-slits. 

It should be pointed out that this opening stage of the 
larval development appears to be of the nature of a rest- 
ing phase, during which the larva accumulates energy for 
future growth. 

Pr&oral "Nephridium" of Hatschek. 

In the larvae of Amphioxus there is a structure lying at 
the base of the notochord on the left side, immediately 
above the praeoral pit, which we have not yet consid- 
ered. (Cf . Figs. 8 1 and 82, ;r.) According to Hatschek, who 
first described it, it arises in the larva as a mesodermal, 
ciliated funnel and canal in front of the mouth, in the 
region of the first metamere. It lies in a narrow division or 
prolongation of the body-cavity, beneath the left aorta. (Cf. 
Fig. 76 B.} At its hinder end it opens into the pharynx. 
Hatschek interprets this structure as a nephridium. Its 
true physiological, and especially its morphological, sig- 
nificance is, however, very perplexing and requires further 


Ancestral Number of Gill-slits. 

The unlimited number of gill-slits in the adult Amphi- 
oxus has led to a good deal of controversy as to the ap- 
proximate number present in the ancestral Vertebrate, 
some authorities being of the opinion that Amphioxus 
presents the primitive condition in this respect, and 
others that the multiplication of gill-slits in this form 
was a secondary phenomenon. 

Sometimes as many as fourteen pairs of gill-clefts are 
found in a remarkable cyclostome fish from the Pacific, 
allied to Myxine, and called Bdellostoma* With this ex- 
ception, no true fishes, recent or fossil, are known which 
possess more gill-slits than the existing sharks belonging 
to the family of the Notidanidce. Of these the genus 
Hcptanchus possesses eight gill-clefts (i.e. seven plus the 
spiracle) on each side, and Hexanchus seven. In Ammo- 
ccetes, as we have seen, there are at one time indications 
of eight pairs of gill-slits. The first pair of these, how- 
ever, never breaks through to the exterior, and eventually 
disappears, but Dohrn has shown that the primary rela- 
tion in which the seventh pair of cranial nerves stands to 
it, indicates that it is the homologue of the spiracle of the 
higher forms. 

Moreover, in the larval development of Amphioxus 
several facts combine to produce the impression that the 
indefinite number of gill-slits in the adult is a secondary 
acquirement. First of all, there is the series of primary 
gill-slits which, while varying within narrow limits, usually 
numbers fourteen. Their unpaired unilateral character is 
merely incidental, as explained above, and it may be stated 

* For a recent account of Bdellostoma, consult HOWARD AYERS, No. 69, 


that they are potentially paired, the first of them in all 
probability being actually paired (with the club-shaped 

In the second place, after the closure of a number of 
the primary slits, the so-called critical stage occurs with 
eight pairs of gill-slits. This is another resting phase in 
the development, and marks the turning-point from the 
larval to the adolescent period. Subsequently the addi- 
tion of new gill-slits behind those already present com- 
mences and goes on indefinitely throughout life. 

Counting in the first pair of slits (i.e. first primary slit 
plus club-shaped gland) which is destined to atrophy, we 
must regard it as probable that the proximate common 
ancestor of Amphioxus and the higher Vertebrates was 
characterised by the presence of from nine to fourteen 
pairs of gill-clefts, although it is also probable that there 
was a variable tendency to add to this number by fresh 


1. (p. 105.) It is unaccountable how there can have been 
conflicting statements as to the ejection of the genital products 
(male and female) through the atriopore. It was first observed by 
DE QUATREFAGES in 1845, an( ^ ms observations have since been 
fully confirmed by PAUL BERT, A. WILLEY, and E. B. WILSON. On 
the other hand, both KOWALEVSKY and HATSCHEK affirm that they 
are discharged through the mouth. It is to be regretted that two 
such eminent observers should have committed this error, since it 
is difficult to eradicate it from the text-books. 

2. (p. 115.) The primitive endoderm cells in the neighbour- 
hood of the neurenteric canal apparently retain an undifferentiated 
character, until the completion of the myotome-formation. In the 
young embryo they are to be observed in transverse section in pro- 
cess of division, numbers of karyokinetic figures being present. 
But the cells divide without regard to the median plane of sym- 

NOTES. 1/5 

metry, and the recent researches of E. B. WILSON and LWOFF lead 
to the conclusion that the so-called mesoblastic pole-cells, which 
were described by HATSCHEK, have no real independent existence. 

3. (p. 123.) Whether the dorsal and ventral fin-spaces are 
actually derived from the original myoccel, as described by Hat- 
schek, or do not rather arise by a splitting of an originally solid 
thickening of the gelatinous connective tissue which surrounds 
them, must remain doubtful. The cavity of the metapleural folds 
certainly arises as a schizocwl, i.e. by a hollowing out of a solid 
thickening. Even in case the fin-spaces also arise as schizoccels, 
Hatschek's interpretation of their morphological significance might 
still hold good. 

4. (p. 123.) A transitory pouch-like diverticulum of the myo- 
ccel has been observed in connexion with the formation of the 
sclerotome in the Selachian embryo by RABL and H. E. ZIEGLER. 

5. (p. 129.) Since the work of BALFOUR on the development 
of Elasmobranch fishes (Selachians), it has been known that the 
paired praemandibular head-cavities communicate with one another 
across the median line in the embryo. The important results 
obtained by the researches of KUPFFER (Petromyzon, Acipenser), 
KASTSCHENKO (Selachian), and JULIA PLATT (Selachian), not only 
established the fact that the prsemandibular cavities arose essen- 
tially as anterior archenteric pouches (cf. Fig. 72), but also that 
the median cavity which effected their communication across the 
middle line, from side to side, arose by constriction from the front 
end of the archenteron (using the latter term with some latitude), 
and that, therefore, the union of the right and left prcemandibular 
cavities in the embrvo of the craniate Vertebrates is primary, and 
not secondary, as was previously supposed. 

For an excellent historical and critical summary of our knowl- 
edge of the origin of the head-cavities in the craniate Vertebrates, 
the reader may consult FRORIEP. (See bibliography.) 

6. (p. 130.) The ciliation of the ectoderm in the larva of 
Amphioxus continuing, as it does, long after the muscles have been 
fully differentiated, and when the cilia are therefore no longer 
required for purposes of locomotion, should be especially noted as 
evidence of a very archaic organisation. 

We shall find in the last chapter that the possession of a ciliated 


ectoderm is a prime characteristic of Balanoglossus and many of 
the lower worms (e.g. Nemertines). In none of the craniate 
Vertebrates is the ectoderm at any time ciliated. 

7. (p. 134.) The exact stage at which the club-shaped gland 
reopens into the pharynx must remain an open question. It is, 
very probably, subject to a good deal of variation in this respect, 
occurring now earlier, now later. Experiments to determine the 
physiological role of this gland are much needed. 

8. (p. 143.) In accordance with Dohrn's conception of the 
principle of the change of function (Das Princip des Functions- 
wechseli), the number and nature of the organs of the Vertebrate 
body, which have been interpreted as modified gill-clefts, are truly 
astonishing. First and foremost, Dohrn supposed that the Verte- 
brate mouth arose by the fusion of two gill-slits across the middle 
line, the old Annelid-mouth, which perforated the central nervous 
system, having been lost. A great many forcible arguments have 
been brought forward in support of this hypothesis. Dohrn him- 
self would probably admit that it is only tenable on his further 
hypothesis that Amphioxus is a form which has undergone a retro- 
gressive evolution from the craniate Vertebrates. This was a 
better hypothesis than that of Semper, who, perceiving that 
Amphioxus would not fall in with the Annelid-theory, declared, 
" er sei kein Wirbelthier ; also, auch kein Fisch." 

Besides the mouth, many other structures have similarly been 
referred back to modified gill-slits, among which may be mentioned 
the nose, hypophysis, thyroid gland, lens of the eye, and the anus. 
None of these comparisons is supported by the facts of develop- 
ment and anatomy of either Amphioxus or the Tunicates, while 
most of them would appear to be definitely disproved by these 

9. (p. 147.) Since the right metapleural fold bends round to 
the median ventral line of the snout, as shown in Fig. 38, and 
since, further, at a later period, the right half of the oral hood is 
similarly continued round the front end of the body into the 
dorsal fin, it is clear that the right half of the oral hood must 
arise essentially in continuity with the right metapleur. On the 
contrary, the left half of the oral hood arises entirely independently 
of the left metapleur. It is possible that this discontinuity of 

NOTES. !77 

development of the left half of the oral hood and the left meta- 
pleur has been secondarily brought about. 

10. (p. 150.) The study of transverse sections has led me to 
the conclusion that the backward extension of the endostyle is 
effected by interstitial growth, and not by the conversion of the 
cells which form the primary floor of the pharynx into endostylar 
elements. These cells are probably disintegrated and absorbed 
by the endostyle as it grows backward. 

11. (p. 153.) For a comparison between the perigonadial 
cavities of Amphioxus and the mesonephric tubules of the 
craniates the reader should consult Boveri's original memoirs. 
(See bibliography.) 

12. (p. 159.) The following definition of the so-called bio- 
genetic law of recapitulation (Haeckel's biogentisches Grund- 
gesetz) will explain the meaning of Haeckel's terms " cenogenesis " 
and " palingenesis." According to this law : The development of 
the individual (ontogeny} is a compressed summary of the gradual 
modifications which have resulted in the evolution of the species, 
or type (phytogeny = Stammesgeschichte) ; this recapitulation 
(summary, or Auszug) of the phylogenetic stages in the ontogeny 
is the more perfect according as the ancestral development 
(Palingenesis, Auszugsentwicklung) has been the less disturbed 
or falsified through secondary or " recent " adaptation (ceno- 
genesis, Storungsentwickelung) of the embryo or larva to a new 

13. (p. 162.) The explanation of the asymmetry of the larva 
of Amphioxus given in the text was first suggested by me in 1891. 
It may be well to state that it has not as yet received very general 
recognition in the more recent literature on the subject. It was, 
however, fortunate enough to receive the endorsement of the late 
Professor MILNES MARSHALL in his text-book of Vertebrate Em- 
bryology. When the pelagic larvae of Amphioxus are confined in 
glass jars, after a certain lapse of time they sink to the bottom, 
like all other pelagic organisms. When they arrive at the bottom, 
they fall over on to one side, owing to a physical impossibility to 
rest in any other position, just as was described above for the 
adult. It ought not to require to be emphasised that their inci- 
dentally lying on one side is not due to a pressing desire or 


instinct to assume that position, but rather because they cannot 
help it. It is apparently in consequence of a misunderstanding 
of this observation that KORSCHELT and HEIDER ascribe the larval 
asymmetry of Amphioxus to the same causes which brought about 
the asymmetry of the Pleuronectidae. Another, and, as it appears, 
a still more impossible view, has recently been expressed by VAN 
WIJHE. According to van Wijhe, the left-sided mouth occupies its 
normal and primitive position in the larva of Amphioxus, and in 
that position it represents a gill-slit, whose antimere is the club- 
shaped gland. Van Wijhe arrived at this view as a result of his 
very important discoveries as to the musculature and innervation 
of the adult mouth. These discoveries may be summarised as 
follows : 

1. The outer muscle of the oral hood represents the anterior 
continuation of the left half only of the transverse and subatrial 

2. The inner nerve-plexus of the oral hood is formed on both 
sides, exclusively from nerves which arise from the left side of the 
central nervous system. 

3. The velum is innervated entirely from nerves of the left side. 

From these observations van Wijhe concludes that the mouth of 
Amphioxus, even in the adult, is essentially an organ of the left 
side, and is neither homologous with the Ascidian nor with the 
craniate mouth. 

It would seem, however, that the more obvious and justifiable 
interpretation of these facts is that the asymmetrical musculature 
and innervation described by van Wijhe are merely the partial 
persistence in the adult of the more complete asymmetry of the 

Van Wijhe's observations, therefore, do not affect the question 
of the cause of the asymmetry in any degree. 

14. (p. 165.) As first shown by Dohrn, the hypophysis of 
Ammoccetes first arises from the roof of the stomodceum, from 
which it is subsequently removed to the dorsal surface of the head 
by the enormous development of the upper lip. 

15. (p. 169.) The ciliated tracts in the pharynx of Ammo- 
ccetes were first described and figured by ANTON SCHNEIDER in 

NOTES. 179 

1879. In 1886 DOHRN thought he had proved that the anterior 
portion of them, which bends upwards on either side of the 
pharynx and forms the peripharyngeal grooves, represented the 
last traces of the aborted first pair of gill-clefts. Although they 
appear at the place which was formerly occupied by these rudi- 
mentary gill-pouches, yet, according to Dohrn's own account, they 
do not appear until after the gill-pouches have completely flattened 
out. Under these circumstances, but above all, in view of the 
relations of the homologous peripharyngeal bands in Amphioxus 
which exist both before and after the disappearance of the first 
pair of gill-clefts (i.e. first primary gill-Cleft and club-shaped 
gland), it must be assumed that Dohrn's interpretation, though 
most natural, was nevertheless somewhat at fault. 



THE Ascidians, Tunicates, or sea-squirts, as they are 
indifferently called, constitute one of the most clearly 
defined and yet most heterogeneous groups of animals 
which it is possible to imagine. There is a great variety 
of families, genera, and species occurring all the world 
over, and in all depths of the ocean from the tide-marks 
to the profoundest depths. 

Most of them are sedentary animals, remaining fixed 
all their lifetime on one spot, whether attached to rocks, 
stones, shells, or sea-weeds, from which they are incapable 
of moving. There are, however, several very extraordi- 
nary genera of Ascidians which swim or float about per- 
petually in the open ocean, and have become adapted in 
the extremest manner to a purely pelagic environment. 
These pelagic Ascidians have become so modified in adap- 
tation to their oceanic existence, and their development 
diverges, as a rule, so much from the normal, that they 
will hardly enter at all into the present discussion, with 
the exception of one family, the Appendicularia. 

Just as there are two kinds of sessile Ascidians, simple 
and compound or colonial, so there are two analogous kinds 
of pelagic Ascidians. In some of the latter, however, 
where there is an alternation of generations, one genera- 
tion, namely, the asexual generation, is a solitary form, 
while the sexual generation is a colonial form, as, for 
example, the solitary Salpa and the chain-Salpa. 

i So 


For convenience, the Ascidians, as a whole, may be 
arranged as follows : 



e.g. Ascidia. 



e.g. Appcndicularia. 


e.g. Clavelina. 



(or capable of producing a colony 

by budding) . 
e.g. Pyrosoina. 



The compound sessile Ascidians consist of colonies of 
individuals or ascidiosooids produced by budding from a 
parent individual. Such colonies are often brilliantly 
coloured and of massive proportions, as Amaroucintn and 
Fragariiim. Others form thin encrusting expansions on the 
surfaces of marine plants and shells, as Botryllus and Lepto- 
clinum. In others, again, the individuals are entirely 
separate, except at the base, where they are connected 
together by a common creeping stolon from which new 
buds are periodically produced, as Clavelina and Perophora. 

Test, Mantle, Atrium, Branchial Sac. 

The simple or solitary Ascidians which do not produce 
buds, present hardly less striking differences among the 
different families than do the compound, but their general 
shape is much more uniform. 


An average simple Ascidian, as Phalliisia or Cynthia, 
has been aptly compared to a leather bottle provided with 
two spouts. The spouts occur in the form of two funnel- 
like prominences projecting from the surface of the body 
and bearing at their free extremities the incurrent or buc- 
cal and cxcurrent or cloacal apertures respectively, the 
latter usually occurring at a lower level than the former. 

The most prominent and, apart from the two apertures, 
the only external feature of a simple Ascidian, is the char- 
acteristic tunic or test which surrounds the whole body. As 
a rule, all Ascidian s of whatever kind possess this external 
tunic, and it is one of their chief diagnostic characters. 

According to the species this test may be of a cartilagi- 
nous, coriaceous, fibrous, or membranous consistency, 
usually opaque, but sometimes hyaline and transparent, as 
in Corella, Salpa, etc. Its outer surface may be smooth, 
wrinkled, or rough, capillated, papillated, or mammillated. 
In 1845 KARL SCHMIDT made the discovery that the test 
of the Ascidians was largely composed of the substance 
which forms the cell-walls in plant tissues ; namely, cellu- 
lose. When treated with the proper chemical reagents, it 
gives the cellulose-reaction. This is interesting as show- 
ing the fundamental identity of protoplasm whether it 
occurs in animal- or in plant-cells, since in both cases it 
is capable of depositing cellulose. 

Judging by external appearances an ordinary Ascidian 
resembles nothing so little as Amphioxus, and yet it is 
probably more closely related to the latter than is the 
lamprey larva, Ammoccetes, whose external resemblance 
to Amphioxus is incomparably greater. 

It is only in its internal organisation that we meet with 
structures which remind us strongly of corresponding 
parts in Amphioxus. 


A schematic representation of a dissection of a typical 
Ascidian after Professor W. A. HERDMAN, whose reports on 
the Ascidians collected during the voyage of H. M. S. 
Challenger have done so much to advance our knowledge 
of the group, is given in Fig. 94. The greater part of the 
thick cartilaginoid test (also called tunic, outer mantle, or 
cellulose mantle), t, is supposed to be removed from the 
right side, and its cut edge can be traced all the way round. 
Below the test comes the inner or muscular mantle, w, 
which is the true body-wall, to which the external tunic is 
secondarily superadded. 1 The muscular mantle is limited 
externally (below the test) by the epidermis, and beneath 
the latter are the interlacing muscle-fibres which compose 
the bulk of the mantle. 

Beneath the mantle is an extensive cavity surrounding to 
a large extent the viscera. This is the peribranchial or 
atrial cavity which communicates with the exterior by the 
atrial or cloacal aperture, at.s. 

The mouth, or.s, leads into ft\& pharynx or branchial sac, 
/>//, which is of surprising dimensions, and stretches nearly 
to the posterior end of the body. The walls of the bran- 
chial sac are perforated by innumerable gill-openings, the 
so-called stigmata, arranged in successive transverse rows, 
through which the water which enters at the mouth passes 
out of the sac into the atrial cavity. 

Dorsal Lamina, Endostyle, and P eripharyngeal Band. 

On cutting through its right wall we open into the 
cavity of the branchial sac along the dorsal side of which 
a fold is seen projecting freely into the cavity, the so-called 
dorsal lamina corresponding to the dorsal groove in the 
pharynx of Amphioxus, while along its ventral side is a 

1 84 







Fig. 94. Diagram of a dissection of Ascidia, from the right side. (After 

The peribranchial cavity is indicated by the black shading. 

an. Anus. at.s. Atrial siphon, e.g. Cerebral ganglion, beneath which is the 
subneural gland and its duct. d.l. Dorsal lamina, end. Endostyle. g. Gonad. 
g.d. Genital duct. int. Intestine, m. Muscular mantle, oes. Aperture, leading 
from branchial sac into oesophagus, or.s. Buccal siphon, ph. Branchial sac. 
st. Stomach, t. Test or cellulose mantle, tn. Buccal or coronary tentacles. 
ty. Typhlosole ; internal fold of intestinal wall, to increase the digestive surface. 


well-defined groove with white glistening walls, which is 
the endostyle. The groove of the endostyle is deeper here 
than in Amphioxus, but its epithelial walls have the same 
histological differentiation, with the two rows of gland- 
cells on each side of the middle line, the latter being 
occupied by a median group of cells carrying very long 
cilia. The food which enters the mouth together with the 
water does not pass out of the pharynx into the atrial 
chamber, but is caught up by the slime secreted by the 
endostyle and is then carried fonvards along the endostyle, 
and, having arrived at the anterior extremity of the latter 
at the base of the buccal tube, is carried round along a 
circular ciliated groove which surrounds the base of the 
mouth at the entrance to the branchial sac, until it reaches 
the dorsal side of the animal, when it is led backwards by 
the ciliary action of the cells of the dorsal lamina in the 
form of a cord of slime in which the food-particles (micro- 
scopic organisms, vegetable debris} are imbedded. 

The ciliated groove round the base of the buccal tube 
connecting the anterior extremity of the endostyle with the 
dorsal lamina is known as the peripharyngeal band or 
pericoronal groove. We have already made the acquaint- 
ance of the homologue of this structure both in Amphi- 
oxus and in Ammocoetes. It forms a complete circle 
round the base of the buccal tube and is indicated in 
Fig. 94 by the black line which limits the pharyngeal. 
wall anteriorly. It is still better shown in Fig. 96, which 
represents a young individual of Clavelina. 

The cord of slime containing the food passes backwards 
along the dorsal lamina to the opening of the oesophagus, 
which lies near the posterior end of the branchial sac, in 
the dorsal middle line, through which it passes into the 
stomach. The dorsal lamina is continued to one side of 

1 86 


the oesophageal aperture, as a low ridge, which joins the 
posterior extremity of the endostyle.* 

Visceral Anatomy. 

Except in its most anterior region, the dorsal border of 
the pharynx lies freely in the atrial chamber. On the 
contrary, along its ventral border, throughout the whole 

length of the endo- 
style, it is attached to 
the muscular mantle. 
In other words, the 
right and left halves 
of the atrial cavity are 
continuous round the 
dorsal side of the 
pharynx, but are 
separated from one 
another ventrally by 
the concrescence of 
the endostyle with 
the mantle. (Cf. Fig. 
95.) In Amphioxus, 
as we have seen, the 
opposite condition ob- 
tains. There, the dor- 
sal wall of the pharynx 

95- Diagrammatic transverse section is closely applied to 
through the middle of the body of Ascidia. (After , , , , , 

HERDMAN.) The muscular mantle is indicated ' d > wn 

by the black shading. 

a. Peribranchial cavity traversed by numerous 
vascular trabeculae, through which the blood flows 
into the branchial bars. br.s. Branchial sac. 

b.v. " Blood-vessels." d.l. Dorsal lamina, e. Endo- r , ... , . .. 

of the ciliated tracts in the 
style. ec. Ectoderm, g. Gonad. gd. Double 

genital duct. /'. Intestine, with typhlosole. r. Rec- 
tum, r.o. Renal vesicles. /. Test. 

the endostylar tract 

* Compare the above with 
description of the course 

pharynx of Ammoccetes, 
given on p. 1 68. 


is free, so that the right and left halves of the atrial 
cavity are continuous ventrally, instead of dorsally. 

In order to see the stomach and intestine, it is necessary 
to cut through the left wall of the pharynx, since the vis- 
cera lie, at least in the genus Ascidia (or Phallusia), on 
the left side of the pharynx. It should be pointed out 
that the topographical arrangements vary considerably 
among the different genera of Tunicates. In Clavelina, 
for example, the viscera lie behind the pharynx, as shown 
in Fig. 96. 

On the left side of the pharynx (Fig. 94) the short 
cesophagus leads into the dilated stomach, which again 
narrows down to the looped intestine, and finally the lat- 
ter bends sharply forwards into the rectum, which opens 
by the anus into the atrial cavity, the excrement being 
carried to the exterior by the constant stream of water 
which flows out through the atrial or cloacal aperture. 

Instead of being straight, as in Amphioxus, the aliment- 
ary canal is here doubled round upon itself. This U-shaped 
character of the alimentary canal of Ascidians is shown 
with great clearness in the case of Clavelina (Fig. 96), 
where there are no secondary convolutions in the course 
of the intestine. 

The Ascidians are one and all hermaphrodite, and the 
reproductive glands frequently lie between the loops of 
the intestine, while two ducts, oviduct and vas deferens, 
which often present the appearance of a single duct with 
a double lumen, proceed forwards by the side of the rec- 
tum, to open into the cloacal region of the atrial cavity 
near the anus (Fig. 94,^- and gd\ 

The ovary and testis, though quite separate in the adult, 
originate, according to the account given by the Belgian 




from a common centre of formation, which subsequently 
undergoes a division into two portions, one of which be- 
comes the ovary, and the other the testis. Similarly the 
oviduct and vas deferens are derived by division of a 
primarily single structure, which arises in continuity 
with, and in fact as an outgrowth from, the primitive 
sexual gland. 

In spite of their hermaphroditism, it would appear that 
not all the Ascidians are self-fertilising, although many, if 
not most of them, are. In some cases it is supposed that 
in different individuals the male and female organs attain 
maturity at different times, so that in a given individual, 
when the ovary is ripe the testis is unripe, so that it must 
be fertilised from another individual, in which the testis is 
ripe, but the ovary unripe, and so on. 

Nervous System and Hypophysis. 

{Neurohypophysial System . ) 

The central nervous system of an Ascidian usually bears 
f a ridiculously small proportion to the bulk of the organ- 
ism. Its main constituent is a ganglion which lies im- 
bedded in the thickness of the mantle, between the oral 
and the atrial siphons, the two latter structures being 
innervated by nerves proceeding from the ganglion. As 
belonging to the central nervous system must also be 
mentioned a solid nerve-cord which runs along the dorsal 
border of the branchial sac from the cerebral ganglion 
to the visceral region (Fig. 96). This was discovered by 
van Beneden and Julin, and is derived from a persistent 
portion of the central nervous system of the larva. 

Beneath the cerebral ganglion is a lobulated glandular 
organ known as the subneural gland. It is provided with 



a duct which runs forward and opens at the end of a 

ciliated funnel-shaped dilatation into the branchial sac at 

the base of the buccal tube 

(Figs. 94, 96, and 97) in 

front of the peripharyngeal 


The branchial opening of 
the duct of the subneural 
gland appears primarily as 
a simple circular orifice, but 
it does not usually retain 
this character in the adult. 

Generally it assumes a 
crescentic form by the in- 
curving of its anterior or 
posterior lip, and then in 
many cases the horns of the 
crescent so formed become 
coiled over and over con- 
centrically, and usually in 

, u Fig. 96. Young Clavelina, shortly 

approximately the Same after the metamorphosis, from the right 

plane, so that the lips of side - (After VAN BENEDEN and JULIN.) 

at. Atrial opening, at.c. Atnal cav- 

the aperture assume a Very ity. b.s. Blood-sinus, end. Endostyle. 
, . -. ep. Epicardium ; outgrowth from bran- 

complicated appearance and chial sac behind e ndost y ie, which grows 
constitute the so-called dor- down into the cree P' n g stolon - formin s 

a septum in the latter, and being the 

Sal tubercle (Fig. 97). chief element in the production of buds. 

,, . . / Lobes of the fixing organ, which give 

It has taken a long time rise to the creep ing stolon, g. Ganglion. 

and the WOrk Of a great f-*- Stigmata. A. Heart, hy. Hypophysis 

. (dorsal tubercle), int. Intestine. m. 

many Zoologists tO achieve Mouth, oes. (Esophagus. p.b. Periphar- 

i T yneeal band. pc. Pericardium. /. Re- 

our present knowledge ma s ins of tai]> 4 hdrawn into the body. 

(which is by no means * Visceral nerve. 

complete) of the subneural gland of Ascidians and its 

1 90 


Fig- 97- Hypophysis of Phallusia 
mentula, prepared out and seen from the 
inside. (After JULIN.) 

g. Subneural gland, above which may 
be seen the outline of the ganglion and its 
nerves, d. Duct of the subneural gland. 
t. Dorsal tubercle, the opening of the 
hypophysis into the branchial sac. The 
actual opening is indicated in black. 
pc. Peripharyngeal groove. ep. Epi- 
branchial groove, d.l. Dorsal lamina, 
slightly displaced, to show the duct of the 
subneural gland above it. 

N.B. In this species, the atrial and 
buccal siphons are widely separated, and 
the duct of the subneural gland is very long. 

The dorsal tubercle was 
discovered by the celebrated 
SAVIGNY in 1816, and was 
for a long time supposed to 
be an independent sense- 
organ of an olfactory nature. 
The subneural gland was 
detected not as a gland, but 
as an enigmatical structure 
lying below the brain by 
the English naturalist HAN- 
COCK in 1868. Its glandular 
character was demonstrated 
by NASSONOFF and Ussow 
in 1874-75, the last-named 
author showing its connex- 
ion by means of the duct 
with the dorsal tubercle. In 
1 88 1 JULIN produced an 
admirable memoir on the 
subneural gland and its duct, 
and strongly urged its ho- 
mology with the pituitary 
body or hypophysis cerebri 
of the higher Vertebrates. 
The same suggestion was 
made in a more tentative 
form in the same year by 
BALFOUR. We shall have 
to consider this question 
later. Suffice it to say at 
present that Julin's sugges- 
tion has been accepted to 


the extent that the subneural organ of the Ascidians is 
frequently spoken of as the hypophysis. 

Circulatory System. 

With regard to the circulatory system the Ascidians 
differ markedly from Amphioxus in the possession of a 
well-defined heart which lies in a distinct pericardium. 
The heart lies ventrally and usually in the neighbourhood 
of the stomach. (Cf. Fig. 96.) Its wall is muscular, but 
consists only of a single layer of cells whose deeper portions 
(i.e. towards the cavity of the heart) are drawn out into 
striated muscular fibres, while the outer portions of the 
cells containing the nuclei project into the cavity of the 

There is therefore no true endothelial lining to the heart, 
and the cells which build up its wall offer a most interest- 
ing example of epithelio-muscular tissue, as was first pointed 
out by Edouard van Beneden. This type of muscular tis- 
sue, in which the muscle-fibres occur as basal prolonga- 
tions of cells which still retain their epithelial character, is 
found, as is well known, in the case of the body-muscles of 
the Nematode or thread-worms, and is above all character- 
istic of the Ccelenterata (Hydroids and Medusae). 

There are no true blood-vessels in Ascidians, but the 
passages along which the blood percolates are merely 
lacunae in the connective tissue and musculature of the 
body and between the viscera. They are not lined by an 
endothelium, and are more correctly described as. blood- 
sinuses. They are often irregular in their outline, as shown 
in the transverse section represented in Fig. 95, but often 
again they simulate the appearance of true blood-vessels, 
as in the case of those branches which pass from the 
mantle into the substance of the test, as well as the tubes 


which traverse the wall of the branchial sac in every 

In the second chapter it was pointed out that the 
Vertebrate heart arose as a specialisation of a portion of 
the primitive sub-intestinal blood-vessel whose calibre was 
originally uniform throughout, and that in Amphioxus the 
cardiac region of the vascular system retains its primitive 
tubular character. 

Very different is the actual origin of the Ascidian heart ; 
although it is simply a dilated tubular structure, yet it 
arises entirely independently of and prior to the rest of 
the vascular system at a time, in fact, before the formation 
of the muscular mantle and before the atrial cavity has so 
far extended itself as to almost entirely replace the original 
body-cavity. The blood-sinuses of the Ascidians are rem- 
nants of the latter. 

With the formation and growth of the atrial cavity, the 
perforation of the stigmata, and the development of the 
muscular mantle, the original body-cavity becomes reduced 
to a system of narrow canal-like spaces which constitute 
the above-mentioned blood-sinuses. The general distribu- 
tion of the blood-sinuses can be made out from Fig. 95. 
There are two main longitudinal sinuses, one below the 
endostyle and another above the dorsal lamina, while 
others are scattered irregularly in the muscular mantle ; 
others again lie in amongst the viscera forming the inter- 
spaces between the various parts ; and finally the bran- 
chial bars between the stigmata are all hollow, and their 
cavities are placed in communication with the system of 
sinuses at intervals as shown in Fig. 95. 

The periodic contraction of the heart of Ascidians takes 
place on a highly characteristic and unique plan. Each 
systole occurs as a peristaltic wave of contraction passing 


from one end of the heart to the other ; but the chief 
peculiarity in connexion with it is, that after a certain 
number of contractions in one direction the heart makes a 
brief pause and then commences to contract again in the 
opposite direction, and so it goes on contracting now in one 
direction and now in the other. This phenomenon of 
the periodic reversal of the direction of contraction of the 
Tunicate heart is known as the recurrent action of the 
heart, and was discovered in 1824 by VAN HASSELT. 
The discovery was first made in the case of Salpa, but it 
has since been found to hold good for all Tunicates. 

When the heart contracts from its posterior to its an- 
terior extremity, that is to say, in the postero-anterior 
direction, the blood is thereby propelled forwards into the 
blood-sinus which lies below the endostyle, and from this it 
passes into sinuses which run transversely into the bran- 
chial bars. In the basket-work formed by the intercross- 
ing of the branchial bars, the blood has a complicated 
and irregular course, and is finally collected into the dorsal 
sinus which lies above the dorsal lamina. Here it flows 
backwards, and after passing in amongst the viscera arrives 
back to the heart. (Other branches of the sinuses pass 
into the test, where they end in curious knob-like dilata- 

On the contrary, when the heart contracts in the reversed 
or antero-posterior direction, the blood which has already 
been oxygenated in its passage through the branchial bars 
is sent to the viscera direct, and from there it collects 
into the dorsal sinus, from which it is distributed over the 
branchial sac, and so into the sub-endostylar or ventral 
sinus, in which, it flows backwards to the heart. 

On account of the above peculiarities relating to its 
independent origin, the histological structure of its wall, 


and its recurrent action, the Tunicate heart would appear 
to be a unique organ peculiar to the group of the Ascid- 
ians and analogous but not homologous, or only incom- 
pletely so, with the heart of the Vertebrates. 

Again, the vascular system of an Ascidian is only func- 
tionally comparable to that of Amphioxus, since true vessels 
provided with an endothelial lining are entirely absent, 
their place being taken by sinuses which arose by reduction 
from the original body-cavity. 

Renal Organs. 

The renal organs of the Ascidians have no apparent 
morphological relation to those of Amphioxus, and therefore 
need not detain us. They consist of a group of bladder-like 
vesicles with cellular walls lying around the intestine. The 
products of excretion (uric acid, etc.) are deposited inside 
the vesicles in the form of solid concretions. There is no 
excretory duct. In Molgula, there is a single large cylin- 
drical renal sac closed at both ends and lying on the right 
side of the body, behind the heart, known as the organ of 

Comparison between an Ascidian and Amphioxus. 

Having sketched in rough outline the organisation of an 
adult Ascidian, we are now in a position to consider in 
what respects it resembles and in what it differs from that 
of Amphioxus. We shall see that some of the most funda- 
mental differences will be made good by the structure of 
the larva, such as the absence of a dorsal nerve-tube and 
of a notochord. 

Let us first consider the resemblances between an adult 
Ascidian and Amphioxus. 


In both cases the pharynx is perforated by a great num- 
ber of gill-apertures (gill-slits, stigmata), converting it into 
a branchial sac and opening into an atrial or peribrancliial 
cavity instead of directly to the exterior. At the base of 
the pharynx there is a longitudinal gland consisting of a 
groove open throughout its whole length towards the cavity 
of the pharynx, and known as the endostyle, whose histo- 
logical character is closely similar in the two cases. From 
the anterior extremity of the endostyle a ciliated band of 
columnar cells passes round the wall of the pharynx on 
each side, in front of the gill-openings, and abuts on the dor- 
sal border of the pharynx, along which it is continued back- 
wards in connexion with the dorsal lamina in the one case 
and the hyperpharyngeal groove in the other. This band 
forms a circlet round the pharynx behind the velum, and is 
the peripJiaryngcal band* We shall find also that the 
Ascidian hypophysis is essentially homologous with the 
olfactory pit of Amphioxus. 

In the Ascidians there are sphincter muscles round the 
buccal and atrial siphons, and inside the former, in front of 
the peripharyngeal band (pericoronal groove), there is a 
circlet of tentacles corresponding perhaps to the velar 
tentacles of Amphioxus. (Cf. Fig. 94, /.) 

The differences between the structure of an adult Ascid- 
ian and of Amphioxus may appear to outweigh the resem- 
blances, but it must be remembered that they are all 
correlated with and accessory to the one great difference 
in the mode of existence of the respective types. 

An Ascidian is sessile ; Amphioxus is free. The former, 
as it were, builds its house upon a rock and is immovable ; 
the latter lives in the shifting sands, and is capable of 
extremely active locomotion. 

* As mentioned above, this band is usually grooved in the Ascidians. 


In correlation with this sessile habit of existence we find 
that the Ascidians, in contrast to Amphioxus, are hermaph- 
rodite, an almost universal condition among sessile organ- 
isms of every description. They are unsegmented, the 
muscles not being divided up into myotomes ; and none of 
their organs (gonads, renal organs, etc.) are metamerically 
repeated, unless we regard the successive transverse rows 
of sti;mata in the wall of the branchial sac as evidence of 


metamerism. It is, however, of a totally different nature 
from the metamerism of the gill-slits of Amphioxus, and 
we shall see that only in the earlier stages of their devel- 
opment can the stigmata of the Ascidians be compared 
with the former. 

Another of the most characteristic accompaniments of 
a sessile mode of life is the U-shaped alimentary canal. 
Instead of being a straight tube with a posteriorly directed 
anus as in Amphioxus, the alimentary canal of the Ascid- 
ians is doubled up upon itself, the rectum is directed for- 
wards, and the anus opens into the atrial cavity. The 
absence of a dorsal nerve-tube and notochord in the adult 
Ascidian has been indicated above. 

In spite of these great differences, the presence of the 
endostyle and the perforated wall of the pharynx in the 
adult, and above all the features in the embryonic and 
larval development, entitle the Ascidians to be defined as 
more or less Amphioxus-like creatures which have become 
adapted to a sessile habit of existence. 


The first accurate and detailed account of the embryonic 
development of Ascidians was the classical memoir pub- 
lished in 1867 by KOWALEVSKY in the Memoires de 
1' Academic imperiale des Sciences de St. Petersbourg. 


The Ascidian larva was known long before this time, 
and the external features of its metamorphosis were de- 
scribed in 1828 jointly by AUDOUIN and MILNE-EDWARDS, 
to whom the discovery of the free-swimming larva is due. 
Furthermore, the internal structure of the tailed larva, 
and even the histological structure of the axial rod of the 
tail, was described with some accuracy by KROHN in 1852, 
but in ignorance of the details of the embryonic devel- 
opment, he was unable to give the right morphological 
interpretation to the various parts, and did not identify 
the axial rod with the notochord of the higher forms. 

Segmentation and Gastrnlation. 

The segmentation of the egg, the formation of a hollow 
one-cell-layered blastula, and the flattening and subse- 
quent invagination of one side of the blastula to form 
the two-cell-layered gastrula, take place on a plan so 
essentially similar to what has been described above for 
Amphioxus that it is not necessary to dwell at length 
upon them here. Suffice it to point out that the segmen- 
tation of the Ascidian egg takes place typically, according 
to VAN BENEDEN and JULIN, on a strictly bilateral plan. 
That is to say, when the ovum has divided into two 
blastomeres, right and left, each blastomere represents 
and will give rise to the corresponding half of the larval 
body, and the descendants of the first two blastomeres 
can be distinguished for a remarkably long time on each 
side of the middle line of the embryo, a fact which is 
highly characteristic of Ascidian development. 

After the gastrula has begun to elongate, and the blas- 
topore has been narrowed down by the approximation of 
its lips to a small aperture situated at the posterior dorsal 
extremity of the embryo, the formation of the medullary 
plate occurs. 


Formation of Medullary Tube and Notochord. 

Here, as in Amphioxus, the dorsal wall of the embryo 
flattens, while the ventral remains convex, and the ecto- 
dermic cells on the dorsal side become marked off from 
the rest by their larger size and columnar shape. The 
medullary plate extends nearly to the front end of the 
embryo, while posteriorly its cells form a ring round 
the blastopore. 

In the formation of the medullary tube, however, there 
is an important difference, and the Ascidian embryo con- 
forms in this point more to the mode of development 


f ~en 

^/^^ I I I ^^J A n "-C rt 1 -^ X-^l^ 


Pig. 98. Transverse sections through embryo of Clavelina Rissoana, to show 
mode of formation of medullary tube and mesoderm. (After DAVIDOFF.) 

A. Through anterior region of embryo, with medullary groove still open. 

B. Through posterior region, with closed medullary tube. 

ch. Rudiment of notochord. ec. Ectoderm, en. Endoderm. mes. Mesoderm. 
m.g. Medullary groove, m.t. Medullary tube. 

which is typical of the higher Vertebrates than does 
Amphioxus. In the latter the medullary plate sinks 
bodily below the level of the surrounding ectoderm, which 
then grows over it. Subsequently while underneath the 
ectoderm the medullary plate assumes the form of a 
half-canal open towards the ectoderm, and eventually its 
margins come together and so form a complete tube. 

In the Ascidian embryo the overgrowth of the surround- 
ing ectoderm and the folding up of the margins of the 


I 99 


medullary plate occur simultaneously, so that when the 
latter has the form of a half-canal it is not closed over 
by a layer of ectoderm, but is open to the exterior 
(Fig. 98). 

At a somewhat later stage the two medullary folds meet 
together and fuse in the middle line (Fig. 98 /?), and this, 
combined with a slight forward growth of the posterior 
lip of the blastopore, leads to the inclusion of the latter 
in the medullary tube, 
so that we arrive at the 
condition already de- 
scribed for Amphioxus, 
in which the nerve-tube 
opens in front to the 
exterior by the ncnroporc 
and behind into the ar- x^s^tfv^oy 4 

Xo^^irQOf/ .tf 

chenteron by the blasto- 

Fig. 99. A. Embryo of Phallus in mam- 

pore, which has now millata seen in optical section from above, to 

T show notochord. 

become converted into ^ Section through . tail of olde r embryo 

the neiirenteric Canal. of Phallasia mammillata. (After KOWALEV- 


Meanwhile the Cells ch. Notochord. . Ectoderm, en. Endo- 
forming the dorsal Wall derm " ""' MeSoderm ' ' Medullary tube. 

of the archenteron in its posterior two-thirds begin to 
gather themselves together to form the notochord (Figs. 
98 and 99). The cells forming the notochord are at first 
arranged end to end (Fig. 99), and subsequently interlace 
in the manner described above for Amphioxus. 

Origin of Mesoderm. 

At about the same time in which the formation of the 
medullary tube and notochord is going on, the mesoderm 
begins to put in its appearance, and this is the first event 
in the development in which there is an important clif- 



ference between the Ascidian and Amphioxus. The 
mesoderm in the Ascidian embryo does not arise as a 
series of archenteric pouches, but is produced on each side 
by a solid proliferation of cells from the primitive endoderm 
which lines the archenteric cavity. This solid proliferation 

begins in the middle region 
of the embryo near the an- 
terior limit of the notochord, 
and extends backwards (Figs. 
98 and 100). It takes place 
from the dorso-lateral cells 
of the endoderm, in a posi- 
tion corresponding to that 
at which the mesoblastic 
pouches of Amphioxus grow 
out from the archenteron. 
The mesoderm of the As- 
cidian embryo therefore 
Fig. TOO. Embryo of ciaveiina Ris- agrees with that of the em- 

soana seen from above, to show the re- 
lation of parts. (Simplified after VAN bryo of Amphioxus in being 

BENEDEN and JULIN.) derived f h primit i ve 

np. Neuropore. en. Endoderm. ent.c. 

Enteric cavity. m.t. Medullary tube, endoderm, but differs in be- 

mes. Mesodermic band. ch. Notochord. . 

ec. Ectoderm. ing solid and unsegmented.* 

* For a recent and elaborate discussion of the origin of the mesoderm in 
the Ascidians see VON DAVIDOFF'S Untersuchungen zur Entwicklungsgeschichte 
der Distaplia magnilarva, etc., II. Allgemeine Entw. der Keimblatter. Mitth. 
Zool. Stat. Neapel, IX. 1891. pp. 533-651. 

As shown by van Beneden and Julin in Ciaveiina, the primary mesoderm 
of the Ascidian embryo can be detected at a much earlier stage of development 
than in Amphioxus. 

I have studied the origin of the mesoderm in Cynthia papillosa and found 
that the primary mesoderm cells are to be distinguished, by their poverty in 
food-yolk, from the remaining endoderm, at the commencement of gastrula- 
tion (at the so-called plakula-stagi). They occur in the form of a crescent 
round the posterior margin of the blastopore, and are carried in by the invagi- 
nation, and then increase in number by mitotic division. In Cynthia, these 


We thus have two solid longitudinal mesodermic bands 
inserted between the ectoderm and endoderm. Anteriorly 
the mesodermic bands consist of several layers of cells super- 
imposed one above the other (Fig. 98), but farther back 
they consist of only one layer of cells. Both portions of 
the mesoderm namely, the anterior two- or three-layered 
and the posterior one-layered portions arise in continuity 
with one another, but they have different fates, the former 
eventually breaking up into loose cells which float about 
in the body-cavity and constitute the so-called mesenchymc, 
the latter, on the other hand, becoming converted into the 
musculature of the tail ; whence the former is spoken of 
as the gastral and the latter as the caudal mesoderm. 

Outgrowth of Tail. 

In Amphioxus, at the stage corresponding to that of 
which we have been speaking namely, when the embryo 
has an oval or sub-elliptical shape it bursts through the 
vitelline membrane inside which it has already been rotat- 
ing for some time by means of the cilia of the ectoderm, 
and escapes into the open sea. This is not the case, 
however, with the Ascidian embryo. The latter is never 
ciliated externally, and it remains enclosed within the fol- 
licular membrane throughout the whole of the embryonic 
period of development. 

After the stage in question, the growth in the length of 
the embryo is accompanied by a ventral curvature, owing 
to the confined space in which it is contained. Moreover, 
the increase in length is not due to a simple elongation of 
the entire body of the embryo, as is the case with Amphi- 

primary mesoderm cells appear to give rise almost exclusively to the caudal 
mesoderm, while the gastral mesoderm appears to be added in front by prolifera- 
tion from the primitive endoderm as described above. 




oxus, but it is merely due to the outgrowth of the tail from 
the body of the embryo (Fig. 101). 

The structures involved in the outgrowing tail are the 
dorsal nerve-tube, the notochord, the caudal mesoderm, 
which lies on each side of the notochord, and will give rise 
to the muscles of the tail, and finally a solid cord of endo- 
derm consisting of two rows of cells placed side by side 

below the notochord (Fig. 
99 *) As soon as the tail 
begins to grow out, the neu- 
renteric canal becomes ob- 
literated, and shortly after- 
wards the anterior neuropore 
Fig. 101. -- Embryo of Phaiiitsia closes up temporarily. At a 

mammillata in side view, to show com- . . 

mencing outgrowth of tail. (After later period, as we shall see, 
KOWALEVSKY.) ft re opens ; not, however, to 

ch. Notochord. ec. Ectoderm, en. En- 

doderm. mes. Mesoderm ; the cells in- the exterior, but into the 

dicated by dark outlines, beneath which , . , 

may be seen the notochord and caudal DUCCal tube. 

endoderm. n.p. Neuropore. n.t. Medul- ^ g \\\Q tail grOVVS in 
lary tube. 

length, it becomes coiled 

round about the body of the embryo, attaining two or 
three times the length of the latter. 

The cord of endoderm cells in the tail of the Ascidian 
larva has been supposed to represent a rudimentary intes- 
tine homologous with the straight intestine of Amphioxus, 
the larval tail being on this view equivalent to the 
post-branchial portion of the trunk in Amphioxus. This 
view, however, is probably not correct, although there is 
something to be said in favour of it. The probability is 
that the tail of the Ascidian larva or tadpole, as it is often 
called, is an organ which has been specially elaborated in 
the course of its evolution for the particular benefit of the 
Ascidians, since (exclusive of the pelagic forms) it is their 


sole organ of locomotion, and hence of transportation from 
place to place ; this only being possible during the larval 

As a rule, the larval phase of an Ascidian's existence is 
a remarkably brief one, and there is on this account all the 
more need for an effective propelling organ, which will 
enable the larva to arrive at a suitable resting-place. 

In Amphioxus, as described above, locomotion is ef- 
fected by serpentine movements of the whole trunk in 
virtue of its muscle-segments, and there is therefore no 
need for a tail in addition ; but there is, nevertheless, a 
short post-anal extension of the body, which alone can be 
regarded as the homologue of the tail of the Ascidian larva. 
In the latter (e.g. Ciona, Phallusia, etc.) the muscles are 
entirely confined to the tail, none being formed in the body 
proper, until after the resorption of its caudal appendage. 

On the view which I am endeavouring to make clear, 
it follows that the tail of the Tunicate tadpole is of the 
same nature as that of the Amphibian tadpole, and, in fact, 
of the craniate Vertebrates generally, and, as has just been 
said, is only represented by the short post-anal section of 
the trunk in Amphioxus. 

The solid cord of endoderm in the tail is not, therefore, 
a rudiment of a primitive intestine, but it is analogous to, 
even if not, as first suggested by BALFOUR, homologous 
with, the so-called post-anal gut which occurs in the em- 
bryos of the higher Vertebrates, and bears a similar rela- 
tion to the formation of the tail that the endoderm-cord in 
the Ascidian embryo does. 

Thus in the typical Ascidian embryo the elongation of 
the trunk (body proper) does not take place to any consid- 
erable extent during the embryonic or even larval period, 
but only after the metamorphosis. 


With the formation of the tail the enteric cavity be- 
comes confined as a closed sac to the anterior portion of 
the embryo. It is bounded dorsally by the nerve-tube, 
which is somewhat dilated in this region, and in front, at 
the sides and below, it is in close contiguity with the 

Formation of the Adhesive Papilla. 

At a much later stage than that represented in Fig. 101, 
the ectoderm bounding the convex anterior extremity of 
the body becomes raised up into three prominences, whose 
relations to one another are those of the corners of a tri- 
angle. They are due to the ectodermic cells at the respec- 
tive points assuming a high columnar shape. They become 
eventually raised very much above the adjoining surface of 
the ectoderm, and become the adhesive papilla or fixing 
glands of the larva. The cells composing them acquire the 
power of secreting a viscid substance, by which the larva 
can fix itself to any favourable surface (Fig. 102). 

Cerebral Vesicle and its Sense-organs. 

We have spoken above of the dilated anterior portion of 
the nerve-tube. This is the part of the central nervous 
system which undergoes the most striking subsequent 
changes. By a gradual widening of its cavity, accom- 
panied by a local thinning out of its wall, this portion 
of the neural tube lying in front of the notochord becomes 
transformed into a spacious sub-spherical vesicle, known 
as the cerebral vesicle (Fig. 102). 

While the anterior portion of the neural tube is enlarg- 
ing to form the cerebral vesicle, granules of black pigment 
are deposited by certain cells in the dorsal wall of the 
vesicle. The granules are at first scattered about in the 




interior of the cells. The most anterior of the cells con- 
taining the pigment is at first distinguished from the 
others solely on account 
of the fact that the pig- 
ment-granules which it 
contains are somewhat 
larger than those in the 
succeeding cells. (Cf. 
Fig. 103.) 

Later on, 
the first 
is seen to separate itself 


Fig . I02 _ Embryo of Ascidia mentula 

Cell snort 'y before hatching; from the right side. 
(After WILLEY.) 

ch. Notochord, undergoing vacuolisation. 
r .1 .1 i . f- Eye. ent.c. Enteric cavity, f. Adhesive 

from the others, and it papil ^ ^ Anterior por(io y n O f nerve . tube 

becomes gradually trans- (spinal cord), o. Otocyst, lying on the floor 

of the cerebral vesicle and projecting up 

ferred by a differential freely into its cavity, r.a. Right atrial involu- 
4.1- r 4.V. 11 r tion. st. Stomodoeum. 

growth of the wall of 

the vesicle down the right wall to its final position in the 
ventral wall of the vesicle (Figs. 102, 103). This cell is 
the otocyst, and the pigment-granules become consolidated 
together to form the otolith. The latter is apparently 

Fig. 103. Optical sections through cerebral vesicle of embryos of Ascidia 
mentula, to show mode of origin of eye and otocyst. (After WILLEY.) 
e. Eye. o. Otocyst. 

extruded from the cell (otocyst) in which it was originally 
formed, and the latter assumes a cup-shape, in the hollow 
of which the otolith lies. The two structures together 
form the so-called auditory organ, whose function may be 
not so much of an auditory nature as that of an equilibrat- 
ing apparatus. 


The other pigment-cells of the dorsal wall of the cerebral 
vesicle collect themselves together and form a slight pro- 
tuberance in the right dorso-lateral corner of the vesicle, 
while the pigment-granules, which were at first scattered 
about in the interior of the cells, become concentrated at 
their converging extremities towards the cavity of the 
vesicle. And in this way is formed the single eye of the 
Ascidian tadpole ; the original pigment-producing cells 
constitute the retina, which retains its primitive position 
as part of the epithelial wall of the brain.* 

Subsequently two or three cells from the adjoining wall 
of the vesicle take up a position, one above the other, in 
front of the mass of pigment and, having previously, by 
an alteration in the character of their protoplasmic con- 
tents, acquired a high refractive index, constitute the lens 
of the eye, which projects obliquely downwards into the 
cavity of the vesicle. (Cf. Fig. 105 A.} 

The cerebral vesicle of the Ascidian tadpole is the un- 
doubted homologue of the corresponding, but less pro- 
nounced, structure in Amphioxus. It differs from the 
latter in lying wholly in front of the anterior extremity of 
the notochord, in possessing a more highly organised eye, 
provided with a cellular lens, and in the presence of an 
otocyst, which, as we have seen, is evolved from the same 
group of cells which gave rise to the eye. 

The eye of the Tunicate tadpole agrees fundamentally 
with the type of eye peculiar to the Vertebrates, in that 
the retina is derived from the wall of the brain. On this 

* The fact that the lens of the Tunicate eye as well as the retina and 
the otocyst arise by differentiation of one and the same epithelial layer of 
the primitive cerebral vesicle, has recently been described by SALENSKY for 
the larva of Distaplia, magnilarva. (W. SALENSKY. Morphologische Studien 
an Tunicaten : I. Ueber das Nervensystem der Larven u. Embryonen VOK 
Distaplia magnilarva. Morph. Jahrb. XX. 1893. PP- 4^-74- ) 


account it is called a my e Ionic eye. In the typical Inverte- 
brate eye, on the contrary, the retinal cells are differen- 
tiated from the external ectoderm. 

Comparison of Tunicate Eye with the Pineal Eye. 

The Tunicate eye, however, differs essentially from the 
paired eyes of the craniate Vertebrates in that the lens, as 
well as the retina, is derived from the wall of the brain. 
The lens of the lateral eye of the Vertebrates is derived 
by an invagination of the external ectoderm, which meets 
and fits in with the retinal cup at the end of the optic 

It is, therefore, an extremely interesting fact which was 
pointed out by BALDWIN SPENCER, that the Tunicate eye 
agrees, in respect of the origin of its lens, with the parietal 
or pineal eye. of the Lacertilia, in which the lens is likewise 
derived from cells which form part of the wall of the- 
cerebral outgrowth which gives rise to the pineal body. 

The pineal body is another of those remarkable rudi- 
mentary structures whose constant presence in all groups 
of Vertebrates forms such an eminently characteristic 
feature of their organisation. It develops as a hollow 
median outgrowth from the dorsal wall of the brain 
(thalamencephalon), the distal extremity of which dilates 
into a vesicle and becomes separated from the proximal 

For a long time the pineal body was a persistent enigma 

* According to the most recent work on the subject the distal vesicle be- 
comes entirely constricted off from the primary epiphysial (pineal) outgrowth 
of the brain, and the parietal nerve does not represent the primitive connex- 
ion of the pineal eye with the roof of the brain, but it arises quite inde- 
pendently of the proximal portion of the epiphysis. 

See A. KLINCKOWSTROM, Beitr'dge zur Kcnntniss des Parietulauges. 
Zoologische Jahrbiicher (Anat. Abth.), VII. 1893. pp. 249-280. 


and the subject of much speculation, one of the most cele- 
brated hypotheses with regard to its significance being 
that of DESCARTES, who regarded it as the seat of the 

More recently it has been shown to represent a rudi- 
mentary, unpaired eye. Although in most cases, curiously 
enough, it exhibits in existing forms no trace of an eye- 
structure, it has been shown by DE GRAAF and SPENCER 
that, as a matter of fact, in many lizards the distal vesicle 
does actually become converted into an eye which, though 
of a rudimentary character, is possessed of a retina, pig- 
ment, and lens. In these forms the pineal body pierces 
the roof of the cranium, occasioning the parietal foramen, 
which is so characteristic of the Lacertilian skull, and the 
pineal eye lies outside the cranium immediately below the 
skin, through which it can be distinguished in external 
view by the presence of a modified scale placed above it. 

In the animals below the lizards in the scale of organi- 
sation (Amphibians and Fishes), as well as in those above 
them, the distal vesicle of the pineal body apparently does 
not become so far differentiated as to be recognised as 
an actual eye, except in the case of the Cyclostome fishes, 
where, as shown by BEARD, it presents the three essential 
elements of an eye ; namely, retina, pigment, and lens, 
lying, however, inside the cartilaginous cranium. 

The facts in our possession would seem to indicate 
that the remote ancestors of the Vertebrates possessed 
a median, unpaired, myelonic eye, which was subsequently 
replaced in function by the evolution of the paired eyes. 
It would, however, be premature either to assert this or to 
express it as a definite opinion, especially since, in refer- 
ring to the evolution of the paired eyes of Vertebrates, 
we are bordering on ground upon which I have no imme- 


diate intention of treading. The pineal eye may not have 
been primitively so much an organ of vision as a light- 
perceiving organ, as is no doubt the case with the eye of 
the Tunicate tadpole. 

We may at least conclude that there can be no doubt 
that the Tunicate eye is the functional homologue of the 
pineal eye of the higher Vertebrates, as Spencer sug- 

Stomodceal and A trial Involutions. 

By the time that the cerebral vesicle of the Ascidian 
embryo with its contained sense-organs (eye and otocyst) 
is approaching the completion of its full development, no 
less than three ectodermic invaginations occur in the body 
of the embryo. One of these is situated immediately in 
front of and in contact with the anterior wall of the cere- 
bral vesicle, the blind end of the involution pressing 
against the subjacent endoderm. This is the stomodceum, 
and its formation is preliminary to the perforation of the 
mouth which takes place later, and places the stomodosum 
in open communication with the portion of the enteric 
cavity which will become the branchial sac (Fig. 102). It 
should be emphatically noted that the stomodoeal invagi- 
nation occurs in the dorsal middle line immediately adja- 
cent to the anterior extremity of the central nervous 

The other two ectodermic invaginations occur symmetri- 
cally, one to the right and the other to the left of the 
dorsal middle line, behind the region of the cerebral vesicle, 
and constitute the pair of atrial involutions, which, by their 
subsequent growth and modification, give rise to the atrial 
or peribranchial cavity. We see, therefore, that the epi- 
thelium which forms the lining membrane of this cavity 
is, as in Amphioxus, derived from the external ectoderm. 


For some considerable time after the metamorphosis the 
young Ascidian possesses two separate atrial cavities, right 
and left, each opening to the exterior by its own atrial 
aperture. Eventually the two cavities extend round the 
branchial sac dorsally, so that their walls come into contact 
in the dorsal middle line, and finally the dividing line 
breaks down, and they become continuous one with another 
dorsally, remaining separated ventrally, as described above. 

At the same time that the two atrial cavities grow 
towards one another, their external apertures become in- 
volved in the same process of growth, and, moving together, 
finally fuse in the dorsal middle line, and so form the single 
atrial or cloacal aperture of the adult.* 

Beyond agreeing in its ectodermal origin, there might 
appear to be not much in common between the mode of 
development of the atrial cavity in the Ascidians and in 

No morphologist would recognise a fundamental differ- 
ence in the fact that the right and left halves of the atrial 
cavity in Amphioxus arise by a single median involution of 
the ectoderm, instead of from a pair of involutions, and that 
they are from the first continuous with one another instead 
of becoming so secondarily (Fig. 104). 

In like manner, the fact that the two halves of the atrial 
cavity are continuous with one another ventrally in Amphi- 
oxus and dorsally in the Ascidians, is easily brought into 
correlation with the other differences in the organisation 
of the two types, which have been described above, and is 
no bar to our regarding the atrial cavity of the one as being 
homologous with that of the other. 

* The time at which the atrial cavities fuse together varies very much in 
different genera. In Molgula manhattensis, for instance, whose stigmata 
develop on a similar plan to those of Ciona (see below), there is a single 
atrial aperture at the moment of the metamorphosis. 


One feature in connexion with the formation of the 
atrial cavity, in which the Ascidians stand in marked 
contrast to Amphioxus, does, however, require a special 

Whereas in Amphioxus the atrial involution has the form 
of a longitudinal groove, in the Ascidians it occurs on each 
side, as a local inpushing of the ectoderm with a minute 
circular orifice of imagination. 2 

The fact has already been stated above that the elonga- 
tion of the body proper of an Ascidian embryo or larva does 
not, in the main, take place until after the metamorphosis. 

A B 

Fig. 104. Diagrammatic transverse sections, to illustrate the mode of forma- 
tion of the atrium in (A) an Ascidian and (B) Amphioxus. (After WiLLEY.) 

The atrial involutions occur at a time when the tail is 
rapidly increasing its length ; the body proper, on the con- 
trary, remaining stationary so far as increase in size is 
concerned, and retaining at this stage approximately the 
dimensions which it possessed when the tail first began to 
grow out. Moreover, they occur before the appearance of 
any gill-clefts in the wall of the branchial sac, so that in the 
Ascidians the gill-slits never open directly to the exterior. 

In Amphioxus, on the other hand, there is no such delay 
in the elongation of the body of the embryo, but it goes on 
continuously till the full complement of myotomes has been 


formed. The post-anal portion of the body, which we sup- 
pose to be the homologue of the tail of the Ascidian tad- 
pole, does not appear until a somewhat late period in the 
development. There is very little of it present in the larva 
with three gill-slits (Fig. 73). 

The reason of this, as explained above, is that the post- 
anal section of the trunk is of only minor functional sig- 
nificance in Amphioxus, but is all-important to the 
Ascidian larva, and consequently, as is the case with 
many other structures of great functional importance in 
the various groups of the animal kingdom, it exhibits a 
precocious development. 

Not only, therefore, has the elongation of the body of 
Amphioxus already taken place before the occurrence of 
the atrial involution, but the primary gill-slits have also 
broken through the wall of the pharynx, and open freely to 
the exterior before the atrium begins to be closed in. 

In Amphioxus, then, the atrial involution has been drawn 
out into the form of a longitudinal groove because it 
occurs subsequently to the elongation of the body and 
the perforation of the gill-slits. 

In the Ascidian embryo the (paired) atrial involution 
has the form of a simple pit with a circular margin, be- 
cause it arises before the elongation of the body proper 
of the embryo and before the perforation of the gill-clefts, 
so that no influence has been at work to draw it out into 
the form of a groove. 

We see, therefore, that a great many of the differences 
between the Ascidian tadpole and the larva of Amphi- 
oxus can be explained sufficiently to allow of their being 
brought into genetic relation with one another, by consid- 
ering the relative time at which corresponding develop- 
mental processes take place in the two cases. 



The following table will help to make this matter 











Oval embryo with medullary 
tube, neurenteric canal, 
notochord, and mesoblast. 
(Last two commencing.) 

Outgrowth of tail. 

Continued growth of tail. 

Formation of stomodosum and 
atrial involutions. 

Escape from vitelline mem- 

Commencing perforation of 

Metamorphosis and commenc- 
ing elongation of body 



Oval embryo with medullary 
tube, neurenteric canal, 
notochord, and mesoblast. 
(Last two commencing.) 

Commencing elongation of 
body of embryo, and escape 
from vitelline membrane. 

Continued elongation of em- 

Formation of mouth, and com- 
mencing perforation of gill- 

Continued formation of gill- 
clefts and outgrowth of tail 
(i.e. post-anal section of 

Formation of longitudinal atrial 


Of course the above table has no concern with the 
actual time (hours and days) from the commencement 
of the development at which such and such an event 
occurs. The type of Ascidian referred to in the above 
description is a simple Ascidian like Ciona or Phallusia. 

The above table also shows how the development of 
the Ascidian and of Amphioxus moves along parallel 
lines up to a certain point, and then at the time of the 
outgrowth of the tail in the embryo of the former and the 
hatching of the embryo of the latter, divergences set in. 


It has long been recognised that the development of an 
Ascidian is much abbreviated in comparison with that of 
Amphioxus, since in the former it neither comes to the 
formation of a ciliated embryo nor to the production of 
archenteric pouches for the mesoderm. One of the chief 
evidences, however, of abbreviation in the Ascidian devel- 
opment is the precocious formation of the larval tail. 

Formation of Alimentary Canal and Hatching of Larva. 

When the enteric cavity of the Ascidian embryo begins 
to grow in length so as to give rise to the stomach and 
intestine, which it does shortly after the appearance of 
the atrial involutions, there is only one resource open to 
it on account of the limited space in which it lies, and that 
is to double round upon itself. This it accordingly does. 
As the growth progresses, the posterior dorsal angle of 
the enteric cavity bends sharply downwards on the right 
side, and then upwards and slightly forwards on the left 
side, ending at first blindly in the vicinity of the left atrial 
sac. In this way the four divisions of the alimentary 
canal become established ; namely, pharynx or branchial 
sac, oesophagus, stomach, and intestine. (Cf. Fig. 105.) 

By the time these changes have taken place, the embry- 
onic development is at an end, and the larva is ready to 
hatch. By spasmodic jerkings of its tail, the larva finally 
succeeds in bursting the egg-follicle or vitelline membrane 
in which it has been hitherto enclosed, and so escapes 
into the open sea. 

Clavelina and Ciona. 

While the development of most forms of Tunicata is re- 
ducible to a common type, yet the details vary within very 
wide limits in different genera. The tendency here, as 


elsewhere, is to abbreviate the development by omitting 
certain ontogenetic processes, and so arriving at the de- 
sired end, as it were, by a short cut. 

One of the most impressive instances of such an abbre- 
viated development, and one which can be demonstrated 
with the utmost certainty, is afforded by the genus Clave- 
lina, in contrasting it with the closely allied genus dona. 

Clavelina (see Fig. 96) is an Ascidian, provided at its 
base with creeping processes or stolons containing a lumen 
continuous with the body-cavity, by which it adheres to 
rocks and weeds. Buds are formed from the stolon, which 
grow up into new individuals precisely like the parent form 
which developed from the egg, and so a colony is produced. 

Ciona also has similar basal processes of the test, con- 
taining prolongations of the original body-cavity, but no 
buds are produced. 

In Clavelina, the embryonic development, up to the time 
of the hatching of the larva, takes place inside the peri- 
branchial chamber of the parent, which becomes converted 
into a kind of brood-pouch. 

In Ciona, the eggs are extruded into the water, where 
they are fertilised by the simultaneous extrusion of sper- 
matozoa from the same individual. Finally, in Clavelina 
the egg is much larger and contains more food-yolk than 
that of Ciona. 

We see, therefore, that in these two genera the egg is at 
the outset subjected to different sets of conditions, both 
internally and externally. 


Three stages in the metamorphosis of the larva of Ciona 
intestinalis are shown in Fig. 105. First, there is the free- 
swimming larva, which, after a pelagic existence of one or 


perhaps two days' duration, is on the point of fixing itself 
to a foreign object by means of the sticky secretion of its 
three adhering papillae. 

This larva possesses features which we have not yet 
considered. Let us give our attention in the first place 
to the tail. 

Vacuolisation of the Notochord. 

The vacuolisation of the notochordal tissue, which was 
described above for Amphioxus, has already proceeded to 
such an extent that there is no longer any trace of cellu- 
lar structure in the centre of the notochord. It is entirely 
filled with a perfectly colourless substance, probably of 
gelatinous consistency, while the nuclei have been dis- 
placed entirely from the centre and can be seen to lie 
closely pressed against the dorsal and ventral sides of the 
sheathing membrane of the notochord (Fig. 105 A). 

There is one respect in which the above vacuolisation 
of the cells of the notochord differs considerably from 
the corresponding process in Amphioxus and the higher 

Whereas in the latter forms the vacuoles appear inside 
the individual cells, in other words, are intraccllnlar, in 
the Ascidian tadpole they occur between the cells, and are 
therefore intercellular. This was first made out by 
Kowalevsky, and can readily be observed. (Cf. Fig. 102.) 
The intercellular spaces separate the cells which were 
previously fitted accurately together, end to end, and, 
gradually increasing in size, they eventually flow together 
and so constitute a continuous space, while the cells with 
their nuclei become thrust aside. 

Assuming that the vacuoles contain a more or less fluid 
substance secreted by the protoplasm of the cells, the 


above difference in the vacuolisation of the notochordal 
tissue in Amphioxus and the Ascidian larva would resolve 
itself into saying that the secretion was retained inside 
the cells in the one case, and deposited outside them in 
the other. 

Mesenchyme and Body-cavity. 

The endoderm cells of the tail, which formed at first a 
solid cord below the notochord, have now become con- 
verted into loose corpuscles, which have mostly floated 
out of the tail into the hinder portion of the body-cavity, 
and have become indistinguishable from the mesoderm- 
cells. The latter are beginning to lose their compact dis- 
position in the form of the two mesodermic bands, espe- 
cially in the hinder region, and to be scattered about in the 

The body-cavity of the young Ascidian is not unre- 
servedly homologous with that of Amphioxus, on account 
of this remarkable behaviour of the mesoderm. The 
cavity does not arise in the midst of the mesoderm by a 
splitting apart of its component cells, but it is simply 
produced by a separation of the endoderm from the ecto- 
derm, the two layers being at first in contact at the sides 
and below ; in fact, everywhere, except where the dorsal 
nerve-tube intervenes. 

In the cavity thus produced between ectoderm and 
endoderm the mesodermic bands at first lie freely, and 
then their component cells break away from their compact 
association and float about the cavity in the form of 
scattered corpuscles, known collectively as mesenchyme. 

This mesenchyme later gives origin to the muscula- 
ture of the body proper of the Ascidian, and also to 
the definitive blood-corpuscles, genital organs, and renal 


vesicles.* All these structures are differentiated from the 
loose mesenchyme cells, all of which at first course round 
about the body of the young Ascidian like blood, being 
kept in motion by the beating of the heart. 

In the stage shown in Fig. 105 A the mesodermic bands 
are still fairly compact in front, having extended them- 
selves anteriorly at the sides of the enteron by interstitial 

PrtEoral Body-cavity and Praoral Lobe. 

When the larva first hatches, the endoderm and ecto- 
derm are in contact with one another at the anterior 
extremity of the body, just as they are in the earlier 
stages. (Cf. Fig. 102.) Soon, however, the ectoderm, 
with the adhering papillae, springs away from the endo- 
derm at this point, leaving a space into which the two 
lateral mesodermic bands force their way. 

In this way a special anterior portion of the body-cavity, 
praeoral and praeenteric, is produced, and is at first com- 
pletely filled by a compact mass of rounded cells derived 
from the mesodermic bands. 

The end of the body of the larva at which the adhering 
papillae are placed of course corresponds to the tip of the 
snout in Amphioxus. 

Just as Amphioxus burrows into the sand with its snout, 
so the Ascidian larva fixes itself to the surface of a rock 
or weed by its snout. The anterior or praeoral portion of 
the body-cavity, of which we have just traced the origin, 
is, and subsequently becomes in a still more pronounced 
way, the cavity of the snout, or prceoral lobe. 

* The pericardium arises ventrally from the endodermic wall of the bran- 
chial sac, and the heart is formed by an infolding of the dorsal wall of the 



Fig. 105. Metamorphosis of Ciona intestinalis ; above is represented the 
anterior portion of the free-swimming larva from the left side ; on the left, the 
larva, shortly after fixation, from the right side ; and on the right, the stage at 
which the change of axis commences, from the left side. (After WlLLEY.) 

a. Atrial aperture, b. Branchial sac. ch. Notochord. e. Endostyle. f. Organ 
of fixation, g. Ganglion, h. Neuropore (having reopened into branchial sac). 
i. Intestine. /. Pyloric gland. m. Mouth. . Nerve-tube, oe. CEsophagus. 
&b. Eye. of. Otocyst. /. Pericardium, s. Stomach, st. Stigmata. /. Tail. 


Body-cavity of an Ascidian and C&lom of Amphioxus. 

We must now endeavour to show how the body-cavity 
of the Ascidian can be brought into genetic relationship 
with the coelom of Amphioxus. The question of the 
absence of metamerism in connexion with the origin of 
mesoblast in the Ascidians need not detain us, since it is 
so obviously correlated with their mode of life. It may 
safely be asserted that the Ascidian mesoderm, as a whole,, 
is homologous with that of Amphioxus as a whole, but in 
the details of its origin and fate it is widely different. 

If we figure to ourselves the coelomic epithelium of 
Amphioxus losing its character as a membrane and break- 
ing up into its constituent cells, which would then lie loosely 
in the body-cavity, we should have essentially the same 
condition of things as in the Ascidians. There are numer- 
ous precedents in the animal kingdom for such a disinte- 
gration of an epithelial membrane. 

A most perfect instance of it has been described by DR. 
R. VON ERLANGER* in connexion with the origin of the 
mesoderm in the fresh-water snail, Paludina vivipara. Here 
the mesoderm appears at first in the form of a median 
bilobed archenteric pouch of relatively large dimensions. 
Soon, however, the cells forming the wall of the pouch begin 
to assume irregular shapes, and so disturb the contour of 
the epithelium, and eventually they break apart entirely 
and fill every nook and corner of the available space with 
a loose mesenckyme. Similar out-wanderings of cells from 
an epithelial wall, though not often of such a complete 
character as the instance above cited, are by no means 

* Zur Entwicklung der Paludina vivipara. Parts I. and II. Morpholo- 
gisches Jahrbuch, XVII. 1891. 


A striking example is afforded by the body-cavity 
of the worm-like Balanoglossus, of which we shall speak 

Here, according to BATESON, the cells lining the cavity 
are continually budding off daughter-cells, which fall into 
the cavity, and eventually almost entirely fill it up with 
ntesenchymatous tissue. In this case, therefore, mesen- 
chyme and an epithelial wall coexist. 

Similarly, the epithelial sclerotomc of Amphioxus is rep- 
resented by a mesenchymatous sclerotome in the higher 
Vertebrates. It is not necessary to multiply instances, 
but many others could be adduced. 

If, now, this disintegration of parietal and visceral layers 
of the mesoderm, which we have imagined above to take 
place in the ontogeny of an animal like Amphioxus, be 
supposed to be thrown back in the development, or, in 
other words, abbreviated to such an extent that the pre- 
liminary formation of a continuous coelomic epithelium no 
longer takes place, we should have precisely those condi- 
tions which we actually find in existing Ascidians. 

As in the cases above quoted for purposes of illustra- 
tion, so in the Ascidians the mesenchymatous condition 
undoubtedly originated ancestrally from what we may call 
an epithelial condition. 

In the Ascidians we may conclude, therefore, that while 
ontogenetically the mesenchymatous condition is to all 
intents and purposes primary, from a phylogenetic point 
of view it is pre-eminently secondary or cenogenetic. 

Having made the reservations implied in the above 
statements, we may confidently assert that as a whole 
the body-cavity of the Ascidians is homologous with the 
ccelom of Amphioxus, and we may define the former as 
a ccelom in which the cells, instead of associating together 


to form a lining membrane round the cavity, remain 
independent of one another and scattered about inside the 

Fixation of tJie Ascidian Larva. 

When the larva first fixes itself to some available surface, 
the tail remains for a time stretched straight out and 
almost motionless, giving perhaps an occasional twitch. 
Soon the tail is observed to become shorter and to finally 
disappear, having been drawn within the body proper of 
the young Ascidian. The entire tail, with the whole of 
the notochorcl, musculature, and caudal portion of nerve- 
tube, becomes thus retracted and invaginated into the 
posterior region of the body-cavity, where it forms a coiled 
amorphous mass, which goes through a gradual series of 
histolytic changes, and is finally absorbed by being dissolved 
in the fluid of the body-cavity (Fig. 105 B). 

By the time the tail has been completely drawn up into 
the body, the organ of fixation or snout, as we have called 
it above, becomes drawn out into a long probosciform 
structure in a line with the long axis of the body. Its 
cavity is no longer completely filled with mesoderm-cells 
as it was at first (Fig. 105 A), but it has become so volu- 
minous that its contained cells are loosely scattered about 
(Fig. 105 B). In the concluding chapter we shall endeav- 
our to show, what has been already implied, namely, 
that the organ of fixation is seen to the best possible 
advantage from a morphological point of view in the 
species now under consideration, viz. Ciona intestinalis, 
and that it is homologous with the praeoral lobe (snout) of 
Amphioxus, including under that term both the prasoral 
body-cavity and the praeoral pit, and further that it is 
homologous with the proboscis of Balanoglossus. 


At the stage shown in Fig. 105 A, the lumen of the 
alimentary canal is extremely reduced, and in many places, 
as in the region of the endostyle, e, its opposite walls are 
in actual apposition, so that the lumen at these points is 
almost obliterated. 

This temporary reduction of the lumen of the alimentary 
canal is due to the narrow space into which it has to be 
compressed, combined above all with the relatively enor- 
mous size of the cerebral vesicle, which exercises a great 
pressure on the subjacent dorsal wall of the branchial sac. 
It may be added that the larva of Ciona does not take in 
food independently until after fixation. 

Reopening of Ncnropore ; Degeneration of Cerebral Vesicle; 
Formation of Definitive Ganglion. 

One of the most obvious features of the metamorphosis 
is the rapid expansion undergone by the enteric and body 
cavities and the no less rapid degeneration of the cerebral 
vesicle. This expansion, by relieving the crowded char- 
acter of the various parts, facilitates greatly the study of 
the changes which take place in the internal organisation. 

The neuropore, which we have described above as having 
closed up at an early period, now reopens again and places 
the neural tube that is to say, as much of it as remains 
after the atrophy of the tail --in open communication 
with the base of the buccal tube (Fig. 105 B, ;/). 

The spacious cavity of the cerebral vesicle has vanished, 
and its walls have undergone disintegration, and, except 
for a portion of the dorsal wall which becomes converted 
into another channel, are now represented by a mass of 
histolytic residua filling the original cavity of the vesicle 
and lying below the anterior portion of the nerve-tube. 


This remnant of the cerebral vesicle of the larva with its 
sense-organs becomes eventually absorbed, and the eye and 
otoltth may often be found floating about the body-cavity 
with the ordinary mesenchyme-cells, and occasionally they 
can be seen actually passing through the heart. 

The anterior portion of the nerve-tube itself, which now 
opens into the base of the buccal tube or stomodceum,* is 
derived from a portion of the dorsal wall of the original 
cerebral vesicle which was constricted off from the latter in 
the form of a narrow tube slightly to the left of the mid- 
dorsal line (Fig. 105 B, n). 

Subsequently the cells forming the dorsal wall of this 
portion of the nerve-tube proliferate and form a solid 
thickening which becomes the definitive ganglion of the 
adult (Figs. 105 C, 106, and 107, g). 

The lumen of the nerve-tube behind the region of 
the definitive ganglion finally becomes obliterated by the 
mutual approximation of its constituent cells, and that 
portion of the primitive nerve-tube which in the larva lay 
between the cerebral vesicle and the root of the tail is thus 
represented in the adult by a solid "cordon ganglionnaire 
visceral" (van Beneden and Julin) which starts from the 
posterior end of the adult cerebral ganglion, and, proceed- 
ing along the dorsal side of the pharynx above the dorsal 
lamina, becomes lost among the viscera. (Cf. Figs. 96, 
105, and 107.) 

Below and in front of the definitive ganglion, which 
finally becomes quite separate from the dorsal wall of the 
neural tube, the lumen of the latter persists and becomes 

* According to renewed observations on Ciona, I find that the neuropore 
reopens into the buccal tube precisely in the line of junction of the stomo- 
dceum with the wall of the branchial sac, so that its upper margin is continu- 
ous with the (ectodermic) stomodoeal epithelium, and its lower margin with 
the (endodermic) branchial epithelium. (See below, V.) 



by subsequent extension the lumen of the subneural gland 
and its duct. 

Thus the anterior portion of the primitive neural tube, 
having become constricted off from the cerebral vesicle 
of the larva, and having given rise by proliferation from 
its dorsal wall to the definitive ganglion, becomes bodily 
converted into that structure which we shall call, in agree- 
ment with JULIN, the hypophysis. 

The opening of the latter into the base of the buccal 
tube becomes the dorsal tubercle of the adult. Finally, at 
a much later stage, the glandular portion of the hypophy- 
sis arises by proliferation of 
spongy tissue from the ven- 
tral wall of that portion of 
the newro-hypophysial tube 
which lies immediately be- 
low the ganglion. 

A section through the 
cerebral vesicle of a larva 
of Distaplia, a colony-build- 
ing Ascidian, showing the 
hypophysis in process of 

J r r J Fig. 106. Frontal section through 

being constricted off from cerebral vesicle of a larva of Distapha 

. . . magnilarva, to show the origin of the 

the VeSlCle, IS given in i-lg. ganglion and hypophysis. (After HJORT; 

1 06. In this genUS the COn- combination of two figures.) 

In the larva of Distaplia, the hypophy- 
dition of things generally is sis opens into the branchial sac be- 

,. r hind the stomodoeum. 

very different from what ^ Cerebral vesicle. . Ectoderm. 

Obtains in Ciona, but it is ' Endoderm. g. Ganglion, hy Hy- 

pophysis (neuro-hypophysial tube). 

introduced to show the 

essential similarity in the mode of origin of the hypophy- 
sis in this form, as observed by Dr. JOHAN HJORT. 

In Distaplia, as is also the case to a less extent in 
Clavelina, the ganglion begins to develop from the wall 


of the neuro-hypophysial tube while the latter is still in 
connexion with, and therefore before the atrophy of, the 
cerebral vesicle, thus indicating a hastening in the devel- 
opment as compared with dona. 

The convexity caused in the dorsal wall of the branchial 
sac by the pressure of the cerebral vesicle persists as the 
anterior portion of the dorsal lamina, and in many or most 
simple Ascidians becomes grooved, forming the epi bran- 
chial groove of JULIN (Fig. 97). At present it is merely a 
ridge, the epibrancJiial ridge. 

In Fig. 105 C the proximal (oral) end of the endostyle, 
c, is seen to be connected with the epibranchial ridge by 
the peripharyngeal band, which we have already described 
in the adult. It apparently arises in sitn by simple spe- 
cialisation of the cells forming the epithelial wall of the 
pharynx at this point. 

Primary Topographical Relations and Change of Axis. 

It must be especially noted that the long axis of the 
young Ciona for some time after fixation is identical with 
that of the tailed larva, and therefore the primary topo- 
graphical relations of the various parts are maintained at 
the stage shown in Fig. 105 B, and we can accordingly make 
use of this stage in which different structures are much 
clearer than in the free-swimming larva for the purpose of 
describing the primary topography, which is of the utmost 
importance when it is desired to institute a comparison 
with Amphioxus. 

Since, as we have seen, the details of the embryogenetic 
processes differ in many respects widely from what occurs 
in Amphioxus, we are inevitably compelled to rely to a 
very large extent on topographical relations in order to 
estimate the homology of this or that structure in the 


Ascidians and in Amphioxus. Fortunately there is one 
structure as to whose complete homology, in the Urochorda 
(Tunicata), on the one hand, and the Cephalochorda, on the 
other, no one entertains a doubt, and that is the endostyle. 
We thus have in the endostyle a firm basis upon which to 
ground our deductions. 

In the larva and in the young Ascidian before the 
primary long axis has been disturbed in the way which we 
shall shortly describe, the endostyle is the most anterior 
endodermic structure in the body, and lies dorso-ventrally 
at right angles to the long axis of the body (Fig. 105 A 
and B, c}. 

As described above in the larvas of Amphioxus, particu- 
larly in the younger larvae (see Figs. 64 and 73), the endo- 
style, though lying asymmetrically on the right side, being 
involved in the general asymmetry of the larva, is quite 
anterior in position, in front of all the gill-slits and partly 
in front of, though also partly opposite, the mouth (on 
account of its asymmetry), and almost at right angles (see 
especially Fig. 64) to the long axis of the body. As there 
is only a short stretch of simple endoderm in front of the 
endostyle in the larva of Amphioxus, we may describe it 
as the most anterior differentiated endodermic structure 
in the larva, thus corresponding with remarkable precision 
to the condition described above in the larval and newly 
fixed Ascidian. 

In the middle of the wall of the branchial sac in Fig. 
105 B are seen, somewhat in front of and below the atrial 
aperture, a, of this side, two lens-shaped structures whose 
slightly concave sides face each other. These are the 
borders of the two first-formed primary branchial stigmata 
or gill-clefts. Their actual openings into the atrial chamber 
are at present so small that they can hardly be seen in 



surface-view, but they are situated at the inner or con- 
cave sides of the two thickenings. On either side of the 

latter can be seen the 
ordinary cavity of the 
pharynx proceeding to- 
wards the oesophagus. 
At a later stage the 
openings of the two 
first-formed stigmata 
become distinctly visi- 
ble (Fig. 105 C). Mean- 
while a change of axis 
is taking place in the 
body of the young 

During the extraor- 
dinary change of axis 
which we are about 
to describe the probos- 
ciform praeoral lobe 
(snout, organ of fixa- 
tion) remains station- 
ary, and the rest of the 

Fig. 107. Young Ciona intestinal** after r j ar tiiallv rotates 
the completion of the change of axis ; from the ' ! V 

left side. (After WILLEY.) through an angle of 90 

/, IV. Primary stigmata. a. Anus, situated , 

immediately below the left atrial aperture, end. degrees, USing the Or- 

Endostyle / Organ of fixation ^Ganglion f fixation as a 
Ay. Hypophysis, int. Intestine, Left atrial 

aperture. /.;;/. Longitudinal muscle, m. Mouth, pivot about which it 

oes. (Esophagus. p.b. Peripharyngeal band. _. 

py. Pyloric gland, st. Stomach. /. Coronary turns. In Tig. IO5 6 

tentacles, v.n. Visceral nerve (cordon ganglion- tne rotation which 
naire visceral). 

takes place very gradu- 
ally is only half performed ; while in Fig. 107 it is complete. 
The method of growth by which this rotation takes place 


is of a very singular character, and it is difficult to define 
it in precise terms. 

In this way then the endostyle (and branchial sac 
generally) comes to be placed at right angles to its primary 

Since in Amphioxus the endostyle altered its primary 
axis by a process of independent growth while the long 
axis of the pharynx was constant throughout the develop- 
ment, we find that here again, as in so many previous 
instances, the details by which similar end-results are 
arrived at are widely dissimilar. 

This complete change of axis by which the praeoral lobe 
(organ of fixation) becomes placed at the posterior extrem- 
ity of the body can only be regarded as a cenogenetic 

It is therefore chiefly to the primary relations which the 
various structures bear to one another, before the change 
of axis, that we must turn for purposes of comparison. If 
we do this, we find that the following sequence of organs 
obtains as well in the larva of Amphioxus as in the newly 
fixed larva of Ciona ; namely: i, praeoral lobe; 2, endo- 
style ; 3, mouth ; 4, gill-clefts. 

Formation of Additional Branchial Stigmata. 

After the change of axis of the body, the long axes of 
the stigmata lie transversely. In their further growth 
they go on elongating in the same (transverse) direction, 
and after they have attained a certain size their ventral 
ends that is to say, the ends nearest the endostyle bend 
round towards each other, and from each of the two first- 

* It goes without saying that the primary long axis of the Ascidian larva is 
homologous with the long axis of Amphioxus. 


formed stigmata a minute portion becomes gradually con- 
stricted or nipped off. Thus between and cut off from the 
two original stigmata, there come to lie two intermediate 
stigmata of much smaller size. (Cf. Fig. 107.) 

In this way, then, in Ciona, we arrive at the stage with 
four branchial stigmata on each side of the pharynx. For 
convenience we shall refer to these by the Roman nu- 
merals, I., II., III., and IV. It is a remarkable fact that 
II. and III. do not arise by new perforations, but are cut 
off from I. and IV. respectively. 

On account of the close relations which the two first- 
formed stigmata, I. and IV., bear to one another during 
the production of the intermediate stigmata, their ventral 
extremities coming into contact and apparently some- 
times fusing together so that II. and III. might almost 
be described as a joint production of I. and IV. rather than 
as entirely independent offshoots, one is forced to the 
conclusion that the two first-formed stigmata themselves, 
though they actually appear simultaneously as separate 
perforations, in reality represent the two halves of a 
single primitive gill-slit divided into two by a tongue- 
bar. If, moreover, we examine the exact origin of these 
two stigmata (I. and IV.) by means of transverse and 
horizontal sections, we may become convinced that such 
is indeed the case ; namely, that they represent the two 
halves of a primitive gill-slit which, on account of the 
precocious formation of the tongue-bar between them, 
become perforated separately. 

For the formation of any two or more consecutive gill- 
slits, we usually expect to find separate endodermic pockets 
or pouches of greater or less depth growing out towards 
the ectoderm. (Cf. Figs. 72 and 92.) 

We ought to find something analogous to this in Ciona 


2 3 I 

if the two first-formed stigmata had the value of indepen- 
dent gill-slits. 

Instead, however, of anything approaching to two endo- 
dermic outgrowths, we find at the base of the atrial invo- 
lution a single endodermic ingrowth making its appearance 
(Fig. 108). 

The angles made by this ingrowth with the neighbour- 
ing wall of the branchial sac remain in contact with the 
floor of the atrium, then fuse with it, and finally become 



i"'"<' \rr,y.//':'.,^mt(7jZ^WWik 

Fig. 108. Diagrams illustrating the mode of origin of the two first-formed 
branchial stigmata in Ciona. (After WlLLEY.) 

at. Atrial involution, ec. Ectoderm, en. Endoderm. g.s. Stigmata, t.b. 

perforated (Fig. 108). This is the way in which the stig- 
mata, I. and IV., arise, and it is difficult, if not impossible, 
to interpret the above-mentioned endodermic ingrowth 
otherwise than as a precocious tongue-bar. 

Even in Amphioxus it was seen how the tongue-bars 
of the secondary slits arose relatively much earlier than 
those of the primary slits. If they arose still a trifle 
earlier, we should have the two halves of each slit becom- 
ing separately perforated, just as it happens in Ciona. 
In a species of Balanoglossus an analogous precocious 


formation of tongue-bars, before the perforation of the 
slits, has been described by Professor T. H. MORGAN. 

From what has been said above, we conclude that the 
first four pairs of primary branchial stigmata of Ciona 
(and this probably applies equally to many species of 
Phallusia) represent and are derivatives of one pair of 
primitive, ancestral gill-slits. 

After a comparatively long interval, during which the 
intermediate stigmata, II. and III., increase in length 
transversely, two more stigmata, V. and VI., arise at inter- 
vals, one after the other, by sepa- 
rate perforations behind those 
already formed (Fig. 109). 

On account of the independent 
origin of V. and VI., it might be 
supposed that they would have 
the morphological value of dis- 
tinct gill-slits, and that we had 
before us three pairs of ancestral 




Pig. 109. -Primary branchial gill-slits represented by six pairs 
stigmata of the right side of a o f pr i mary branchial stigmata. 

young Ciona. (After WILLEY.) 

For this interpretation to hold 

good, we should expect to find that in other forms in which 
six primary branchial stigmata were produced, their origin 
was either the same or reducible to the same type as that 
of the branchial stigmata of Ciona. 

This, however, is not the case, since I have found 
that in Molgula manJiattcnsis* a simple Ascidian which 
occurs in great numbers at New Bedford, Mass., the six 
primary stigmata, corresponding precisely to those in 

* My observations on the development of Molgzila manhattensis were 
made at the Marine Biological Laboratory, at Woods Roll, Mass., in the 
summer of 1893. 



Ciona, have a somewhat different mode of origin. The 
two first-formed stigmata (= I. and IV. in Ciona) appear 
simultaneously as in Ciona. Then after growing to a cer- 
tain size, they curve round at their ventral ends, not in 
opposite directions so as to meet each other as they do in 
Ciona, but in the same direction (Fig. no). The recurved 
ends then become constricted off from the parent stig- 
mata. Later on, a fifth gill-opening arises behind the 
first four stigmata by independent perforation, and after 



Fig. no. Diagram illustrating the mode of origin of the six primary bran- 
chial stigmata of Afo/gula manhattensis. The numbers are placed at the ventral 
ends of the slits. The figure is a combination of several hitherto unpublished 
drawings of different stages in the development. /,///, and V arose by separate 

attaining a certain size, it, in its turn, curves round at its 
ventral end, and eventually the sixth stigmatic opening is 
constricted off from the fifth. 

Since the first six primary stigmata have such different 
origins in two different species, it is obvious that in 
attempting to make a comparison with Amphioxus we can 
only use the two first-formed stigmata, because they agree 
in the above-mentioned species, and in many others in 


arising simultaneously, and in representing, in all proba- 
bility, the two halves of a primitive gill-slit, cut in two by 
a tongue-bar. 

The stigmata which are added to these must, therefore, 
be regarded as secondary modifications, hardly comparable 
to the successive formation of new gill-slits in Amphioxus. 

In the Ascidians, therefore, we can only detect the 
representatives of one pair of primitive gill-slits, and there 
is every reason for supposing them to be homologous with 
the first pair of gill-slits in Amphioxus as defined above. 

The six primary stigmata of each side give rise, by re- 
peated subdivision, to the innumerable stigmata of the 
adult, both in Ciona and Molgula. The following de- 
scription, however, applies more particularly to Ciona. 

In the first place, the primary stigmata grow to a sur- 
prising transverse length, and then commence to divide 
into two equal portions by small tongue-like projections, 
which grow across the aperture indifferently from the 
anterior or posterior walls of the respective stigmata, and, 
fusing with the opposite wall, divide the transversely 
elongated slit into two completely separated halves. Then 
each of the latter divides again in the same manner, and 
so the process of subdivision of existing stigmata goes on. 
In this way six transverse rows of stigmata arise. These 
may be distinguished as secondary stigmata, since they 
arise by division from the primary. 

Gradually, by a peculiar process of growth, the long 
axes of the secondary stigmata change their direction, and 
instead of lying transversely they become directed antero- 
posteriorly. This is their definitive position, and the 
stigmata now go on rapidly dividing again, and the num- 
ber of transverse rows of stigmata is in this way doubled, 
trebled, quadrupled, etc., and we thus arrive at the adult 


condition. Out of the multitude of stigmata which are 
present in the adult Ciona only four arise by independent 
perforation ; namely, the primary stigmata I. and IV. 
(which we regard as the two halves of a primitively single 
slit) and V. and VI. 

First Appearance of Musculature. 

By the time the change of axis of the entire body of 
the young Ciona has been effected the musculature 
characteristic of the adult begins to put in an appear- 
ance. In Fig. 107 circular sphincter muscles are present 
round the buccal and atrial apertures. The latter are still 
paired, but are carried by differential growth dorsalwards 
at a later stage, and finally coalesce together in the dorsal 
middle line to produce the single atrial aperture of the 

One strand of the longitudinal muscles of the later 
muscular mantle is likewise to be seen in Fig. 107. It 
tends to branch dichotomously. Posteriorly it is inserted 
on the inner surface of the organ of fixation near the point 
where it joins on to the body. Later new muscle-bands 
arise similar to the first, and become distributed over the 
body-wall in a spreading fan-like fashion, but posteriorly 
they are all inserted in the same region of the organ of 

Alimentary Canal and Pyloric Gland. 

The course of the alimentary canal can be gathered so 
plainly from the accompanying figures (Figs. 105 and 107) 
that it hardly needs a verbal description. From the 
posterior dorsal corner of the branchial sac the oesophagus 
leads into the wide stomach, and from the latter, again, 
the intestine, which often possesses a strangulated appear- 




ance, doubles up obliquely forwards to the left atrial 
chamber, into which it opens by the anus (Fig. 107). 

In the angle made by the outgoing intestine with the 
stojiiach, a blind diverticulum arises. It is at first a sim- 
ple ccecum, but soon begins to branch (Fig. 105 C\ and 
finally forms an arborescent growth embracing the in- 
testine (Fig. 107). This is the so-called pyloric gland, 
and it is probably homologous with the hepatic ccecum of 


It is generally agreed among those who have a voice in 
the matter, that most of the pelagic Ascidians (Salpa, 

Doliolum, Pyrosoma) are 
highly modified forms, spe- 
cially adapted to a pelagic 
life, one of the results of 
which is that their repro- 
duction is marked by a 
complicated alternation of 

It would, therefore, not 
assist us in our comparison 
with Amphioxus to describe 
these types. 

There is, however, one 
family of pelagic Ascidians, 
the Appendicnlaria, with re- 
spect to which there are two 
widely different opinions. 

Fig. in. Appendicularia (Fritii- The Appendicularias are 

/aria) furcata, from the ventral surface. n . . . A 

(After LANKESTER.) pelagic, free-swimming As- 

a.Anus. gl. Unicellular glands, gj. c idianS, whose adult COndi- 

Gill-slits. h. Dorsal hood-like fold of 

integument. ;. Mouth. /. Tail. 

tion is so far similar to the 



larval condition of the fixed Ascidians, that they retain the 
tail as their organ of locomotion throughout life (Fig. 1 1 1). 

The tail is inserted in the middle of the ventral surface 
of the body proper, and is obviously a mere appendage of 
the latter. 

The mouth is terminal or sub-terminal. There is a sin- 
gle pair of branchial stigmata, which open into a pair of 
tubular atrial cavities, whose separate external apertures 
are seen in front, on the ventral surface behind the mouth. 

The alimentary canal is U-shaped, and the anus opens 
on the ventral surface to the right of the middle line, some- 
times behind and some- 
times (according to the 
species) in front of the 
stigmata (Figs, in, 
112). The endostyle 
is always quite anterior 
in position, and some- 
times, as in Fig. 112, 
removed by a consider- 
able interval from the 


Fig. 112. Diagram of the organisation of 
a species of Appendicularia, from the right side. 

In the posterior ex- < After HERDMAN.) 

a. Anus ; the index line was accidentally 

tremity Of the body drawn about % of an inch in front of the anus. 

... b.s. Branchial sac. ch. Notochord. e. Endostyle. 

are placed the gOnadS, ^ Ganglion, from which the nerve-cord proceeds 

male and female, in backwards to the tail, posing to the right of the 

alimentary canal, g.s. Gill-slit, h. Heart, tut. 

close proximity tO One Intestine, m. Mouth. n.c. Nerve-cord, with 

, , . . ganglionic enlargements in the tail. of. Otocyst ; 

another, the tCStlS in beneath which the hypophysis opens into the 

front and the OVarV branchial sac - ov - Oval 7- P- b - Peripharyngeal 

* band. st. Stomach, te. Testis. 

behind. The heart, as 

described by LANKESTER, is a unique example of a func- 
tional organ reduced to the lowest possible level of histo- 
logical structure. It consists simply of two cells placed 


opposite one another and connected together by contractile 
protoplasmic threads, which keep up a pulsating motion. 

The tail is, as might be expected, more elaborately or- 
ganised than that of the Ascidian larva. The dorsal nerve- 
cord is solid, and proceeds backwards from the ganglion, 
passing to the rigJit of the alimentary canal until it reaches 
the tail, along which it is continued, lying to the left of 
the notochord ; it possesses ganglionic enlargements at 
intervals in the tail, from which nerves pass out. 

The caudal musculature also shows somewhat doubtful 
traces of being segmented in correspondence with the 
ganglionic swellings of the nerve-cord. 

In connexion with the cerebral ganglion there is a 
sense-organ in the form of an otocyst, with an enclosed 
otolith, and below this a ciliated pit opens into the ante- 
rior region of the branchial sac, corresponding to the 
hypophysis, or sub-neural organ, of the fixed Ascidians. 

According to one view, Appendicularia is the living rep- 
resentative of the free-swimming ancestor of the Ascidians. 

According to the other view, it is less primitive than the 
fixed Ascidians, and was derived from the latter by the 
gradual increase, from generation to generation, of the du- 
ration of the pelagic existence of the larvae, until they 
ceased to metamorphose, and so retained the larval struct- 
ure throughout life, becoming at the same time sexually 
mature. 3 

These two views are, of course, antagonistic, and the 
former of them is held by a number of well-known author- 
ities. As we are ignorant of the development of Appen- 
dicularia, it is impossible to decide definitely between them. 

With the facts which are at our disposal, however, the 
second view --namely, that the Appendiculariae represent 
Ascidian larvae which have become secondarily adapted to 


a pelagic life, and have acquired the faculty of attaining 
sexual maturity -- would be more in harmony with what 
we know of the relation of Amphioxus to the Ascidians. 
And it would seem that this affinity can be better demon- 
strated through the comparison of Amphioxus, both adult 
and larva, with a fixed Ascidian like Ciona than with 
Appendicularia. 3 

On the latter view, therefore, the so-called metamerism 
of the tail of Appendicularia, on which so much stress has 
been laid, would be simply a secondary elaboration of the 
tail for the purpose of serving as a permanent locomotor 

The dorsal nerve-cord of Appendicularia was regarded 
by FOL as a simple peripheral nerve. We have described 
above how a portion of the primitive nerve-tube in Ciona 
and other Ascidians becomes reduced to a solid nerve. 

It would be of the greatest interest to discover the mode 
of origin of this nerve-cord in Appendicularia. 

Abbreviated Ontogeny of Clave Una. 

In order to demonstrate clearly the relatively primitive 
character of the development of Ciona it is sufficient to 
enumerate a few facts drawn from the development of 
Clavelina as described by Dr. OSWALD SEELIGER. As 
mentioned above, Clavelina is a near relative of Ciona, and 
in the adult condition resembles it very closely in many 

The development of Clavelina was formerly regarded as 
being of a primitive character, but is in reality, more 
especially in the later stages, abbreviated and hastened to 
a remarkable extent. 

Like Ciona it possesses in the adult numerous trans- 
verse rows of stigmata. Each opening, however, arises by 


an independent perforation, so that all those preliminary 
ontogenetic processes which precede the establishment of 
the transverse rows of stigmata in Ciona are dropped out 
of the development of Clavelina.* 

In Clavelina, again, the change of axis of the body 
proper occurs in the unhatched larva ; so does the fusion 
of the two atrial apertures to form the dorsal cloacal 
siphon. The longitudinal muscles of the body proper 
commence to appear in the free-swimming larva, while the 
caudal muscles are enjoying their highest functional 
activity. The vacuolisation of the notochord does not 
proceed so far as in Ciona, since the cells are never actu- 
ally removed from the centre of the notochord, but remain 
as thin discs stretching across the latter, so that the 
vacuolar spaces do not become continuous. 

The behaviour of the organ of fixation in the larva of 
Clavelina is such that it could hardly be recognised as a 

praeoral lobe except in the light of Ciona. 


i. (p. 183.) The test or cellulose mantle of the Ascidians con- 
tains great numbers of cells of various kinds. These were formerly 
supposed to be derived from the subjacent ectoderm of the body- 
wall. KOWALEVSKY has recently shown, however, that the cells of 
the outer (cellulose) mantle of the Ascidians are derived from 
wandering mesenchyme-cells which wander from the body-cavity 
through the ectoderm (either between the ectodermic cells or 
actually passing through the individual cells) into the mantle. 

* A mode of formation of the branchial stigmata, intermediate between 
that of Clavelina and Ciona or Molgula, has been described by GARSTANG 
for Botryllus. In this genus, the primary branchial stigmata all arise by in- 
dependent perforations, and then later become divided up into the transverse 
rows of stigmata. (W. GARSTANG. On the development of the stigmata 
in Ascidians. Proc. Roy. Soc., Vol. LI. 1892.) 

NOTES. 241 

2. (p. 211.) In Clavelina the atrial involutions do not merely 
arise as minute circular invaginations of the ectoderm, but at first 
they appear as short, though quite distinct, longitudinal grooves. 
Compare also the remarkable longitudinal atrial tubes of Pyrosoma. 

3. (p. 238.) There is another possible way of interpreting the 
structure and systematic position of Appendicularia which may 
perhaps be nearer the truth than either of the views mentioned in 
the text. It is not absolutely necessary to suppose that the 
ancestors of Appendicularia were fixed Ascidians ; but both 
Appendicularia and the fixed Ascidians may have descended from 
a common free-swimming stock, and have undergone certain 
modifications in common, such as loss of true vascular system and 
coelom. Then, while the Ascidians proper became adapted to a 
sessile existence, Appendicularia may be supposed to have gone 
to the opposite extreme, and have become adapted to an absolutely 
pelagic existence. In becoming adapted to such a purely pelagic 
or oceanic environment as that of Appendicularia, it is eminently 
conceivable that an animal would have to undergo as radical a 
modification of structure as it would in becoming adapted to a 
sessile existence. (Compare Salpa, Doliolum, etc.) 



" Den Schliissel richtigen Verstandnisses gibt nicht das Hineinpressen 
nener Thatsachen in eine alte Sckablone, sondern das Anfsuchen des 
genetischen Zusammenhangs der Erscheinungen. WEISMANN. 

External Features. 

OF the free-living protochordates, the lowest type of 
organisation is undoubtedly presented by the Enteropneusta 
(Hemichorda), the group to which Balanoglossus belongs. 

Balanoglossus is a remarkable worm-like creature which 
lives buried in the sand or mud of the sea-shore. By 
means of numerous unicellular integumentary glands which 
are distributed over the surface of the body, it secretes a 
mucous substance to which particles of sand adhere, and 
so makes for itself tubes of sand in which it lives at about 
the level of the low tide-mark. It possesses such a 
characteristic external form and odour (like iodoform)as to 
render it peculiarly easy of recognition. 

In front there is a long and extremely sensitive proboscis 
which is capable of great contraction and extension, and is, 
in the living animal, of a brilliant yellow or orange colour. 
Behind the proboscis follows a well-marked collar-region, 




consisting externally of a collar-like expansion of the 
integument, with free anterior and posterior margins over- 
lapping the base of the proboscis in front and the anterior 
portion of the gill-slits behind. 

In the ventral middle line, at the base of the proboscis 
and concealed by the collar, is situated the mouth (Fig. 
113). Following behind the collar is the region of the 
trunk or body proper, which, in the adult of some species, 
reaches a relatively enormous length, even extending to 

Fig. 113. Larva of Balanoglossus Kowalevskii, with five pairs of gill-slits; 
from the right side. (After BATESON.) 

a. Anus. a. p. Temporary pedicle of attachment, c. Collar, ch. Notochord. 
g.s. Gill-slits, m. Mouth, pr. Proboscis. 

two or three feet. The ectodermal covering of the body 
consists in general of ciliated cells, among which are scat- 
tered unicellular mucous glands ; the cilia, however, appear 
to be more prominent on the proboscis than elsewhere. 

In the region of the trunk, which immediately follows 
upon the collar region, there are a great number of paired 


openings on the dorsal side of the body, placing the anterior 
portion of the digestive tract in communication with the 
outer world. These are the gill-slits, and they are arranged 
strictly in consecutive or metameric pairs to the number of 
upwards of fifty in the adult. In their structure, and more 
especially in the possession of tongue-bars, they bear a 
remarkable resemblance to the gill-slits of Amphioxus. 
This is particularly striking in young individuals. As the 
adult form is approached in the development, the bulk of 
the gill-slits sinks below the surface, only opening at the 
latter by small slit-like pores, and thus their true character 
is obscured in a superficial view. 

Projecting into the interior of the proboscis is a rod-like 
structure which arises as an outgrowth from the alimentary 
canal dorsal to the mouth. The lumen of this endodermic 
diverticulum becomes narrowed down and, in fact, partially 
obliterated, while the cells constituting its walls give rise 
to a spongy vacuolar tissue which strongly resembles the 
notochordal tissue of Amphioxus and the higher Verte- 
brates. On account of its dorsal position above the mouth, 
its endodermic origin, and the vacuolisation of its cells, this 
structure was identified by BATESON in 1885 as tne ' noto- 


Nervous System and Gonads. 

The nervous system of Balanoglossus presents many 
features of the utmost interest and suggestiveness. It 
consists essentially of an ectodermal net work of nerve-fibres 
forming the inner layer of the skin (ectoderm) all over the 
body. In this primitive nervous sheath, which envelops 
the whole body, there are certain definite local thickenings. 
Two of these thickenings occur respectively along the 
whole length of the dorsal and ventral middle lines in the 
trunk-region, thus producing the dorsal and ventral median 


longitudinal nerve-cords. In the region of the collar the 
dorsal nerve-cord becomes entirely separated from the 
ectoderm, and this portion of it contains, at least in young 
individuals, a central canal which, from its origin and 
relations, was shown by BATESON, and more recently by 
MORGAN, to be homologous with the central canal of the 
vertebrate spinal cord. Anteriorly the dorsal nerve-cord 
becomes continuous with a specially dense tract of the 
general nerve-plexus at the inner posterior surface of the 


Fig. 114. Diagram of the organisation of Balanoglossus, from the left side. 
(From a drawing kindly lent by Professor T. H. MORGAN.) 

al. Alimentary canal. Ac 1 . Ccelom of proboscis (anterior or prasoral body- 
cavity). 6c' 2 . Coelom of collar. bc z . Coelom of trunk, b.v. Blood-vessel, proceed- 
ing from the so-called heart (which lies at base of proboscis above the noto- 
chord) to the ventral blood-vessel, ch. Notochord. com. Commissure, between 
dorsal and ventral nerve-cords, dn. Dorsal nerve-cord, separated from the integu- 
ment in the collar-region. d.b.v. Dorsal blood-vessel, gl. Proboscis-gland; 
modified coelomic epithelium surrounding heart and front end of notochord. 
m. Mouth, f.v. Pulsating vesicle, lying inside the " heart." v.b.v. Ventral blood- 
vessel, v.n. Ventral nerve-cord. 

proboscis (Fig. 1 14). This proboscidian plexus thins out 
somewhat towards the anterior extremity, but nevertheless 
forms a complete nerve-sheath for the proboscis and indi- 
cates the sensitive character of the latter (Fig. 115). 

The ventral nerve-cord does not extend into the region 
of the collar, but from the point where the collar joins on 
to the trunk the ventral cord is connected with the dorsal 
nerve-cord by a commissure-like thickening of the integu- 
mentary plexus, which passes in the skin on each side 
round the hinder end of the collar-region (Fig. 114). 



Fig. 115. Diagrammatic transverse sec- 
tion through hinder region of proboscis of 
Balanoglossus. (From a drawing kindly 
lent by Professor T. H. MORGAN.) 

D. Dorsal. V. Ventral, be 1 . Proboscis- 
cavity, almost filled up by mesenchymatous 
and muscular tissue,* proliferated from the 
original ccelomic epithelial layer (indicated 
by the black line below the ectoderm). 
p.v. Pulsating vesicle. A. Heart, ch. Noto- 
chord. n.s. Integumentary nerve-plexus. 

The genital organs, 
testes or ovaries, accord- 
ing to the sex of the 
individual, occur as a 
paired metameric series 
of pouch-like bodies or 
gonadic sacs which ex- 
tend backwards far be- 
yond the region of the 
gill-slits. The gonadic 
sacs are suspended in the 
body-cavity by solid cords 
attached to the dorsal 
integument, which be- 
come perforated in the 
spawning season to ad- 
mit of the expulsion of the 
reproductive elements. 


Although there is no muscular metamerism in Balano- 
glossus, yet we have seen that other organs (gill-slits and 
gonads) are arranged metamerically. And in point of 
fact, among those Invertebrates which are not included 
under the phylum of the Articulata, if there is one pecu- 
liarity of organisation more sporadic in its occurrence than 
another, it is metamerism. It may affect the most differ- 
ent organs of the body either collectively or individually, 
and nothing is more patent than the fact that the meta- 
meric repetition of parts has arisen independently over 
and over again in different groups of animals. 1 

* This tissue is not represented in Figs. 114 and 116, although it is present 
throughout the body-cavity. 



Far from assuming as a self-evident fact that the 
extreme metamerism of the Annelids and Arthropods is 
genetically identical with that of the Vertebrates, we have 
every reason to suppose that it has been elaborated entirely 
independently in the two cases, and that the apparent simi- 
larity is due, as already intimated, to a pamllel evolution. 

Body-cavities ; Proboscis-pore ; Collar-pores. 

Corresponding to the 
three regions into which 
the body of Balanoglossus is 
divided, - - namely, probos- 
cis, collar, and trunk, --the 
body-cavity is divided up into 
three systems of cavities. 
These are (a) the anterior 
body-cavity or cavity of the 
proboscis, (/3) a pair of collar- 
cavities, and (j) a pair of 
body-cavities which form the 
unsegmented ccelom of the 
trunk (Figs. 114, 115). 

These cavities arise essen- 
tially as pouches from the 
archenteron (Fig. 117), al- 
though their actual develop- 
ment differs considerably in 
different species (MORGAN). 

The proboscis-cavity is 
placed in communication 
with the exterior by an open- 



Fig. 116. Diagram of the organisa- 
tion of Balanoglossus, from the dorsal 
side. (From a drawing kindly lent by 
Professor T. H. MORGAN.) 

c.p. Collar-pores, go. Gonads. g.s. 
Gill-slits ; the dark lines converging be- 
hind indicate the superficial portions of 
the gill-slits; below the surface are seen 
the free ends of the tongue-bars. //. 
the posterior Proboscis-pore. Other letters as above. 



wall of the proboscis known as the proboscis-pore. In 
B. Kowalevskii this pore lies asymmetrically to the left of 
the dorsal middle line (Fig. 115), while in B. Kupfferi a 
corresponding opening occurs to the right of the middle 

line, so that in this species 
there are two proboscis- 
pores constituting a sym- 
metrical pair. 

The left proboscis-pore 
of Balanoglossus is obvi- 
ously to be compared with 
the praeoral pit of Amphi- 

The collar-cavities also 
open to the exterior by 
pores, one on each side 

underneath the dorsal pos- 
terior free fold of the 

collar, and on a level with 

Fig. 117. Diagrammatic horizontal , , . , , 

section through an embryo of Balanoglos- tfte Opening O f the first 
sus (type of the direct development), to 
show the origin of the body-cavities as 
archenteric pouches. (After BATESON.) 

ap. Tuft of cilia at the apical pole c i_ 

(indication of an apical plate). bc\ Probos- ^PENGEL States that Water 
cis-cavity. *A Collar-cavities. cP. Trunk- j s taken in through the 
cavities, cb. Circular band of cilia. 

collar-pores into the cavity 

of the collar in order to swell the latter up, so that it 
may serve as an accessory organ of locomotion in so far 
as an alternate inflation and collapse of the collar would 
assist the animal in its slow burrowings in the sand. 

These are the 
funnel-shaped collar-pores. 



Alimentary Canal. 

The mouth cannot be closed, as there is no sphincter 
muscle, and accordingly, as the animal progresses through 
the sand, it swallows a large quantity of the latter in 
which food-particles (unicellular organisms, etc.) may also 
be involved. As the sand passes through the intestine, 
it becomes enveloped in the mucous secretion of the intes- 
tinal epithelium, and is ejected through the anus in a cord 
of slime. 

The alimentary canal is a straight tube between mouth 
and anus. In its hinder portion it is usually sacculated, 
i.e. provided with paired 
lateral saccular dilatations 
comparable to the so-called 
intestinal cceca of the Ne- 
mertine worms. (See below.) 
In the region of the pharynx 
the lumen of the alimentary 
canal is incompletely divided 
by lateral constrictions into , *** c 

J Fig. 118. Transverse section through 

tWO portions, an Upper Or the gill-region of Balanoglossus. (After 

, ... . . SPENGEL.) 

branchial portion carrying aL Dige sti V e portion of gut. br. 

the gill-Slits, and a lower Or Branchial portion of gut. 6c9. Third 

body-cavity (trunk ccelom) ; this is also 

digestive portion (Fig. I 1 8). nearly obliterated in the adult by the pro- 
, , , j i liferation of mesenchyme or " paren- 

The latter was compared by chyme ., from its wal dn ^ orsal 
GEGENBAUR * to the endo- nerve - c r d. d.b.v. Dorsal biood-vessei. 

go. Gonad. g.s. Gill- slit. t.b. Tongue- 
Style of the AscidianS, but bar. v.b.v. Ventral blood-vessel, v.n.c. 
, v 11 4.u iU- Ventral nerve-cord. 

it is probable that this com- 
parison, although a very natural and useful one at the time 
at which it was made, will not hold good, since there is 

* CARL GEGENBAUR, Elements of Comparative Anatomy. Translated by 
F. Jeffrey Bell. London, 1878. 


nothing in the structure or development of this part of the 
alimentary tract in Balanoglossus which will bear compari- 
son with the endostyle.* As indicated in the larvae of 
Amphioxus and the Ascidians, it would seem that the 
endostyle first became evolved or differentiated at the 
anterior end of the pharynx, in front of the gill-slits, in 
correlation with the dorsal position of the mouth. 

Development ; the Tornaria Larva. 

The development of Balanoglossus Kowalevskii as made 
known to us by the admirable work of BATESON is what 
is known as a strictly direct development ; that is to say, the 
embryonic, larval, and adult stages follow one another by 
gradual transitions concomitantly with the simple progres- 
sive growth of the individual and without any striking 
metamorphosis. In other species of Balanoglossus the 
larval form is remarkably different from the adult, and 
becomes transformed into the latter by a very distinct 
metamorphosis. The extraordinary larval form here re- 
ferred to was discovered in 1848 by JOHANNES MULLER, 
who named it Tornaria, and regarded it, as did his succes- 
the larva of an Echinoderm (Starfish). 

It was not until 1869 that its true character as the larva 

* A ciliated tract in the floor of the oesophagus of a Tornaria from the 
Pacific has recently been compared to the endostyle by W. E. RITTER. (On 
a New Balanoglossus Larva from the Coast of California and its Possession 
of an Endostyle. Zool. Anz. XVII. 1894. pp. 24-30.) 

The comparison is at present somewhat doubtful. More recently GARSTANG 
has suggested that the endostyle is derived from the adoral ciliated band of the 
Echinoderm larva. (See Fig. 119.) The suggestion is an interesting one, but 
Garstang's idea of the relations of the pneoral lobe is very different to the one 
here set forth. (WALTER GARSTANG, Preliminary Note on a Neiv Theory of 
the Phylogeny of the Chordata. Zool. Anz. XVII. pp. 122-125.) 


of a species of Balanoglossus was demonstrated by ELIAS 
METSCHNIKOFF. Shortly afterwards, Metschnikoff's dis- 
covery was confirmed and amplified by ALEXANDER 

The superficial likeness between Tornaria and such Echi- 
noderm larvae as Bipinnaria or Auricularia is astonishing, 
and a renewed study of the detailed organisation of 
Tornaria, recently made by MORGAN, appears to have 
established the fact, originally insisted upon by Metschni- 
koff, that this resemblance can only be accounted for on 
the ground of genetic affinity. 

In Figs. 119 and 120 two types of larvae, Tornaria 
and Auricularia, are shown side by side ; and although 
unfortunately they are not figured from exactly the same 
aspect, yet it is obvious at a glance that, in spite of certain 
differences which will be enumerated below, they both 
belong to the same category of larval forms. 

A highly characteristic feature of these larvae is the 
remarkable ectodermal ciliated band which constitutes a 
perfectly symmetrical but somewhat complicated undulat- 
ing seam round the body. The larvae are strictly pelagic, 
and swim about in the open sea by means of their cilia ; 
but the latter, instead of being distributed evenly over the 
whole surface of the body, are concentrated in the region 
of the ciliated bands which are composed of thickened 

In Tornaria there are two ciliated bands, viz.: i) the 
above-mentioned undulating seam which is usually known 
as the circnmoral or longitudinal ciliated band, and 2) a 
pastoral circular ciliated band. Only the former is present 
in Auricularia, and the absence of the circular band in this 

form constitutes one of the chief differences between the 


two larvae. 



From a morphological point of view a more striking 
resemblance between the two larvas than that furnished 
by the longitudinal ciliated bands exists in connexion with 
the anterior body-cavity or enteroccel. In the Echinoderm 


Figs. 119 and 120. Auricularia, larva of Synapta (after SEMON) ; and 
Tornaria, larva of Balanoglossus. (After MORGAN.) 

a. Anus. a.p. Apical plate, be 1 . Anterior body-cavity, communicating with 
exterior by the water-pore. 6c 2 , tc 3 . Second and third body-cavities of Tornaria. 
c.b. Circular ciliated band of Tornaria. c.c. Contractile cord between apical plate 
and anterior body-cavity of Tornaria. g.p. Gill-pouches, h.c. Hydrocoel of 
Auricularia (anterior body-cavity), l.c.b. Longitudinal (circumoral) ciliated band. 
I.e. Left enterocoel (body-cavity), m. Mouth, n. Lateral (paired) nerve-band 
of Auricularia. r.e. Right enterocoel. sp. Calcareous spicules. st. Stomach. 
wp. Water-pore. 

N.B. In Auricularia, the margin of the mouth is surrounded by a ciliated 
band discovered by SEMON, and known as the adoral ciliated band. The poste- 
rior, V-shaped portion of this band lies inside on the ventral floor of the larval 

larva this cavity arises as a median pouch of the archen- 
teron, and there is every reason to suppose that it has a 
similar origin in Tornaria, although this point has not yet 


been determined. The primary anterior enterocoel in the 
Echinoderm larva is not quite the same as the correspond- 
ing cavity in Tornaria, since it contains also the elements 
of the general body-cavity. Apart from slight differences, 
the collar-cavities and general body-cavities arise essen- 
tially in the same way in Tornaria as they do in the case 
of the direct developing larva of Balanoglossus (see above).* 
In the Echinoderm larva, however, the paired body- 
cavities do not arise as independent archenteric pouches, 
but they become constricted off from the anterior entero- 
ccel. Making allowance for these deviations in the origin 
of the body-cavities, --deviations which are by no means 
fundamental, since in both cases the body-cavities are 
ultimately reducible to archenteric pouches, --it is an 
extremely striking fact that both in Tornaria and Auricu- 
laria the anterior enterocoel acquires an opening to the 
exterior on the dorsal surface to the left of the middle line. 
This opening is called the water-pore, since it forms the 
outlet (possibly both outlet and inlet) of the water-vascular 
system of the Echinoderm. In Tornaria it persists after 
the metamorphosis as the proboscis-pore, which has been 
described above. 

The Larva of Asterias vnlgaris ; Water-pores and 
Praoral Lobe. 

In view of what was said above as to the occurrence of 
paired proboscis-pores in B. Knpfferi, it is interesting to 
note that sometimes there are two water-pores, a right and 
a left, in Echinoderm larvae. This has been observed by 

* As to the origin of the body-cavities in. different species of Balanoglos- 
sus, MORGAN summarises his observations as follows : " They may arise as 
enteric diverticula, as endodermal proliferations, or even arise from mesenchy- 
matous beginnings." (See MORGAN. No. 125 bibliog.) 





BROOKS and G. W. FIELD in the larvae of a common star- 
fish, Asterias vulgaris. In this case the primary enteroccel 
becomes constricted off from the archenteron in the form 
of two equal pouches. The right and left enteroccelic sacs 
then take up a symmetrical position on each side of the 
larval oesophagus, and each sac next opens to the exterior 
by a water-pore. The pore in connexion with the right 
sac (Fig. 12 1 ) is, however, of a transitory, rudimentary 

character, and soon closes 
up, while the left pore per- 
sists as the definitive water- 
pore. As in Tornaria, so 
here, the cavity of the larval 
body generally, and of the 
praeoral region (prceoral lobe) 
in particular, is the primary 
body-cavity or blastoccel, 
and contains scattered mes- 
enchyme-cells. At a later 
Fig. 121. - Young larva of Asterias stage in the larva of As- 

vulgaris, from the dorsal side. (After ter i as tne right and left 
G. W. FIELD.) 

/./. Praeoral lobe, l.c.b. Circumoral enterOCCelic SaCS, having ill- 
(longitudinal) ciliated band. oes. CEsoph- , , . , . 

agus! r. e . and /.,. Right and left en- creased greatly in length, 

terocoslic sacs, each opening by a " water- meet one another in the 

pore " to the exterior, sf. Stomach, inf. 

Aperture, leading from stomach into in- region of the praeoral lobe 

and fuse together, thus put- 
ting their two cavities into communication across the 
median line. The median portion of the enteroccel thus 
produced extends up into the praeoral lobe, and so the 
primary blastoccelic cavity of the latter is replaced by a 
secondary ingrowth of the enteroccel (Fig. 122). 

Similarly with the metamorphosis of Tornaria, the 
anterior enterocoel, which is at first of very inconsid- 




erable extent (Fig. 120), increases greatly in size, and 
assumes its definite position and proportions as the cavity 
of the praeoral lobe (i.e. proboscis), thus replacing the 
original blastocoelic space, 
while the water-pore remains 
as the proboscis-pore. 

As described in the previ- 
ous chapter, the cavity of 
the praeoral lobe (fixing 
stolon) of the Ascidian tad- 
pole is of the nature of a 
blastocoel or primary body- 
cavity, containing loose mes- 
enchyme-cells, and it is 
therefore of great impor- of J 
tance to note that whether side. (After G. w. FIELD.) 

By a fusion of the two praeoral loops 

the Cavity of the praeoral O f the ciliated band across the apex of the 
, , // 7 praeoral lobe, followed by a separation in 

lobe IS a blastOCCel Or an f he transvers ' e dire ction, the originally 

eiltcrOCCel, the morphological single circumoral band (cf. Figs. 119 and 

121) has become divided into two bands, 

value Of the Structure itself a preeoral ciliated band p.c.b. and a post- 
rpmainQ thp came oral longitudinal ciliated band /.c.. The 

posterior transverse portion of the prae- 
oral ciliated band has undergone a fusion 

Apical Plate of Tornaria. with the front end of the originally dis- 
tinct adoral band (cf. Fig. 119). />./. Prae- 

At the anterior end of oral lobe, into which the enterocoal has 

extended, m. Mouth, r.e. and I.e. Right 

the body, Or, in Other words, and left enteroccelic cavities, st. Stomach. 
... . a. Anus. 

at the apex of the praeoral 

lobe, in Tornaria, there is an ectodermic thickening in 
which nerve-cells and nerve-fibres and a pair of simple 
eyes have become differentiated. This is the so-called 
apical plate, and it constitutes the central nervous system 
of the larva. It can be recognised for some time after the 
metamorphosis at the tip of the proboscis, but eventually 
disappears completely. A similar apical plate occurs in 


a great number of Invertebrate larvae, and is especially 
characteristic of the free-swimming larvaa (Trochophores, 
or Trochospheres) of Annelids and Molluscs. We shall 
return to this later. 

In Tornaria a single contractile cord passes from the 
apical plate to the anterior enterocoel. 

There is no apical plate in Auricularia, nor in most of 
the other Echinoderm larvae ; but there is reason to sup- 
pose that it has been secondarily lost, since a transitory 
ectodermal thickening at the apical pole can frequently 
be observed in the course of their development ; and, 
moreover, in what is probably the most primitive Echino- 
derm larva known (viz. the larva of the Crinoid, Antedoii), 
there is a well-developed apical plate. 

Metamorphosis of Tornaria. 

The metamorphosis of Tornaria, as originally described 
by Alexander Agassiz, takes place with relative sudden- 
ness. According to the more recent account of the meta- 
morphosis given by MORGAN, a marked diminution in size 
occurs ; the internal organs are drawn together in such a 
way that the larval oesophagus, with the gill-pouches (see 
Fig. 120), is drawn backwards into the body, and the 
anterior enterocoel, as already described, is carried for- 
wards into the praeoral lobe. The longitudinal (circum- 
oral) ciliated band, which was the first to develop, is also 
the first to disappear, while the posterior circular band 
persists to a somewhat later stage. 

The Neincrtines. 

It is thus evident that Balanoglossus, especially through 
its Tornaria larva, shows undoubted marks of affinity to 


the Echinoderms. It will next be shown that there are 
certain features in the adult anatomy which apparently 
indicate a distinct genetic relationship to another group of 
the Invertebrates ; namely, the Nemertine worms. 

The Nemertines are elongated, flattened, or cylindrical 
worms, with a smooth ciliated skin and no external seg- 
mentation, occurring, as a rule, in a closely similar habitat 
to that of Balanoglossus, buried in the sand or mud of the 

Like Balanoglossus, they also possess unicellular integu- 
mentary glands, by means of which they secrete a mucous 
substance, to which frequently sand-grains adhere, thus 
producing a tube of sand round the body. Some of them 
reach an enormous length, and one at least must be 
measured in yards (Linens longissimus exceeding three 
yards in length). 

The chief anatomical features which offer material for 
direct comparison between the Nemertines and Balano- 
glossus relate to the ectoderm, proboscis, nervous system, 
mesenchymatous tissue, the reproductive organs, and the 
alimentary canal. 

As for the ectoderm, considered apart from the nervous 
system, it need only be repeated that in both cases it is 
composed of ciliated cells and scattered mucous glands. 

The proboscis of the Nemertines is one of the most 
characteristic organs of this group of animals. It is not 
permanently protruded, and does not serve as an organ of 
locomotion, as in Balanoglossus, but is usually carried 
about entirely withdrawn within the body of the animal, 
from which it can be shot out with great force and rapidity 
when the occasion demands it. During the process of 
extrusion it is turned completely inside out, and conversely, 
during the process of introversion, the retraction takes 

2 5 8 


place from the tip backwards by the in-rolling of its walls. 
According to the graphic description of HUBRECHT, it is 
retracted "in the same way as the tip of a glove finger 
would be if it were pulled backwards by a thread situated 
in the axis and attached to the tip." 

When at rest within the body the proboscis lies freely 
within a hollow cylinder, the wall of which is thick and 
muscular, and constitutes the proboscis-shcat/i (Fig. 123). 



Fig. 123. Diagrammatic transverse section through the middle of the body 
of a Nemertine. (After LANG, Text-book of Comp. Anat.) 

t.m. Basement-membrane, c.m. Circular muscles, d.n. Dorsal or " medullary " 
nerve, d.v. Dorsal blood-vessel, g. Gonads. int. Intestine. Lm. Longitudinal 
muscles, l.n. Lateral nerves, l.v. Lateral blood-vessel. /. Proboscis, p.s. Pro- 

Sometimes beneath the ectodermal epithelium of the 
Nemertine proboscis there is a continuous sheath of nerve- 
fibres, comparable to the nervous plexus in the proboscis 
of Balanoglossus. 

Partly, therefore, on account of its structure, and partly 
on account of its topographical relations when extruded, 
we are led to suppose that a certain homology exists 


between the retractile proboscis of the Nemertines and 
the non-retractile proboscis of Balanoglossus (BATESON). 

In the most primitive Nemertines the nervous system 
consists essentially of a somewhat complicated pair of 
cerebral ganglia and a diffuse nerve-plexus, with nerve- 
cords lying at the base of the ectoderm.* As the cerebral 
ganglia probably belong to the same category as the cere- 
bral ganglia of all other typical Invertebrates, and are not 
represented in Balanoglossus, we can afford to neglect 
them at present. Confining our attention to the ecto- 
dermal nerve-plexus, we find occurring in it, along definite 
lines, local thickenings, after the same principle, but not 
all on the same lines, as was described above for Balano- 
glossus. Directly comparable with the dorsal longitudinal 
nerve-cord of Balanoglossus, there is a similar thickening 
or concentration of the integumentary nerve-plexus in 
some of the Nemertines, in the dorsal middle line (Car- 
inina, Cephalothrix). Hubrecht, who discovered this, calls 
it the medullary nerve. There is, however, no correspond- 
ing ventral nerve-cord in the Nemertines, but, instead of 
this, there is a pair of lateral thickenings, constituting the 
well-known lateral nerves of the Nemertines (Fig. 124). 

It is usually supposed that the lateral nerves of the 
Nemertines are homologous with the two halves of the ven- 
tral nerve-cord in the Annelids. In the Annelids the 
primitive lateral nerves (which are so typical of the Platy- 
helminths, or flat-worms) have approached one another in 
the mid-ventral line, and have often undergone intimate 
fusion together. In some cases, however, they are separated 
from one another by a wide interval (Sabellaria, etc.). 

* HUBRECHT compared the lobes of the cerebral ganglia of a Nemertine to 
the cranial ganglia of the Vertebrates, the lateral nerves to the Kami laterales 
vagi, and the proboscis-sheath to the notochord. 



In the Annelids, in contrast to the Nemertines, the gan- 
glion-cells are not distributed uniformly along the whole 
length of the nerve-cord, but are collected together to 
form definite ganglionic swellings. 

It is, therefore, very significant that in the Nemertines 
we have a median dorsal " medullary " nerve, in addition 
to the elements which constitute the ventral nerve-cord of 
the Annelids. 

In many Nemertines the dorsal and lateral nerve-cords 
do not continue to lie in the ectoderm throughout life, but 


Fig. 124. Diagrammatic view of anterior portion of a Nemertine, from the 
left side. (After HUBRECHT, from LANG.) 

a.l. Anterior lobe of brain. /./. Posterior lobe of brain, p. Opening of pro- 
boscis, m. Mouth, d.n. Dorsal nerve, l.n. Lateral nerve, r.n. Ring-nerves. 

sink deeper into the body, and so come to be separated 
from the ectoderm, first by the basement membrane, and 
then by one or more muscular layers of the body-wall. In 
the Hoplonemertea (those in which the proboscis is armed 
with stylets) the medullary nerve is absent. In all cases, 
however, the longitudinal nerve-cords remain connected 
with one another by a more or less plexiform arrangement 
of nerve-fibres ; although sometimes a more definite con- 
nexion, by means of metameric ring-nerves, has been 
observed by HUBRECHT (Fig. 124). 


There is no true coelom in the Nemertines, and the 
space between the alimentary canal and body-wall is oc- 
cupied by a gelatinous mesenchyme, containing muscular 
and connective tissue elements. In Balanoglossus the cav- 
ity of the ccelom becomes largely obliterated in the adult, 
by the proliferation of cells from the epithelium of its 
walls, thus filling up the cavities with a more or less solid 
parenchymatous tissue. 

Like Balanoglossus, the Nemertines have a straight ali- 
mentary canal, provided with paired lateral outgrowths or 
intestinal coeca, and a terminal anus. 

The gonadic sacs of the Nemertines offer a striking re- 
semblance to those of Balanoglossus. They occur as a 
metameric series of paired sacs, which alternate with the 
above-mentioned intestinal cceca, and communicate with 
the exterior by short tubes, which are at first solid, as in 
Balanoglossus, subsequently becoming hollowed out and 
opening above the lateral cords (Fig. 124). 

Finally it should be pointed out' that, while excretory 
organs, in the form of a well-developed single pair of 
elongated nephridia, provided with numerous internal 
"end-sacs," are present in the Nemertines, nothing of the 
kind has yet been detected in Balanoglossus. 


It is interesting to note that there are some remarkable 
animals which stand in a similar relation to Balanoglossus 
that the Ascidians do to Amphioxus. While Balano- 
glossus is free-living, does not produce buds, and has a 
straight alimentary canal, these creatures, of which only 
two genera are at present known, Cephalodiscus and Rkab- 
doplenra, lead a sessile existence, produce buds, and have 



a U-shaped alimentary canal. Both are deep-sea forms, 
Cephalodiscus having been dredged during the Challenger 
Expedition, from the Straits of Magellan, at a depth of 245 
fathoms ; while Rhabdopleura was first dredged indepen- 
dently, off the Shetland Islands, at 90 fathoms, by the Rev. 

Fig. 125. Cephalodiscus dodecalophus, from the ventral side. (After 

Actual length of polypide from extremity of branchial plumes to the tip of the 
pedicle is about 2 mm. 

b.s. Buccal shield; the shading on its surface indicates pigment-markings. 

At the tip of the pedicle, buds are produced. 

CANON NORMAN, and off the Lofoten Islands, at 200 fath- 
oms, by Professor G. O. SARS (1866-68). Rhabdopleura is 
the name given by ALLMAN (1869), who published a short 
account of it ; and it has since been described by SARS, 


The account which we possess of Cephalodiscus forms 
one of the Challenger Reports, and was written by Pro- 
fessor W. C. M'INTOSH, who made out the main features 
of its anatomy. It was further treated, from a morpholog- 
ical standpoint, by SIDNEY F. HARMER, who pointed out 
its remarkably close affinity to Balanoglossus. 

The most important morphological features in the anat- 
omy of Cephalodiscus are shown in Figs. 125-127. The 
individuals live in colonies, in a " house " or coenoecinm, 
which consists of a ramifying and anastomosing system of 
tubes, the walls of which are composed of a semi-trans- 
parent, gelatinous material, whose outer surface is covered 
with spinous projections. The walls of the ccencecium 
are furthermore perforated by numerous apertures, which 
allow of the ingress and egress of water. 

The adult members of a colony have no organic con- 
nexion between themselves, but each one is independent 
and free to wander about the tunnels of the ccencecium. 
Although Cephalodiscus has not been studied in the living 
condition, there is every reason to suppose that it moves 
about in its tube by means of the large buccal shield (Fig. 
125) overhanging the mouth, by which it can attach itself 
to the inner surface of the tube, and then help itself 
along by the curious pedicle which occurs ventrally near 
the hinder end. It thus seems probable that this pedicle 
can be used as a sucker, but its chief function lies in the 
production of buds which grow out from it, and eventually 
become detached. Bateson has described a somewhat 
similar sucker at the hinder end of the body in young 
individuals of Balanoglossus (Fig. 113). 

Behind and above the buccal shield there is a row of 
twelve tentacles or branchial plumes, each possessing a 
central stem or shaft which carries numerous lateral 




pinnae. An important function of these plumes is to 
produce currents of water by the action of their cilia, 
which vibrate in such a direction that the water with 
food-particles is led into the mouth. The superfluous 
water is led out from the proximal portion of the aliment- 
ary canal by a single pair of gill-slits which are not visible 

in surface view, since they 
are overhung by a fold of 
the integument known as 
the post-oral lamella or 
operculum, corresponding to 
the posterior free fold of 
the collar in Balanoglossus 
(Fig. 126). 

In its internal organisa- 
tion, if due allowance be 
made for its U-shaped ali- 
mentary canal, Cephalodis- 
cus greatly resembles Bala- 
noglossus (Figs. 126, 127). 
The buccal shield of the 
former is obviously the 

Fig. 126. Longitudinal frontal (right 

and left) section through an adult Cephalo- equivalent of the probos- 
discus. (After HARMER.) 

be 1 -. Second portion of body-cavity cis of the latter, and the 
(collar-coelom). bc z . Third portion of . . . , 

body-cavity (trunk coelom). ^-.Pharynx. cavlt Y which it Contains 

c.p. Collar-pores, g.s. Gill-slits, int. in- corresponds to the probos- 

testine. n.s. Nervous system, op. Oper- 

cuium. oes. CEsophagus. st. stomach, cis-cavity. Moreover, the 

t. Base of tentacle. . . _ . . 

proboscis-cavity in Cephalo- 

discus (i.e. the cavity of the buccal shield) communicates 
with the exterior by two proboscis-pores placed right and 
left of the dorsal middle line. 

Following behind the buccal shield is the collar-region, 
from which the branchial plumes arise dorsally, while 



laterally and ventrally it is produced into a free fold to 
form the above-mentioned operculum. The collar-region 
contains a section of the ccelom which is precisely homolo- 



Fig. 127. Longitudinal sagittal section through an adult Cephalocliscus. 
(After HARMER.) 

The section is supposed to be taken sufficiently to one side of the middle line to 
allow of the representation of one of the ovaries and one of the proboscis-pores. 

a. Anus. b.c. Trunk-ccelom. c.c. Collar-ccelom. ch. Notochord. int. Intes- 
tine, m. Mouth. n.s. Nervous system. op. Postoral lamella (operculum). 
ov. Ovary; the oviduct is deeply pigmented. p.c. Praeoral ccelom (cavity of 
buccal shield). ph. Pharynx. /./. Proboscis-pore, pad. Base of pedicle. 
st. Stomach. 

gous with the collar-cavities of Balanoglossus. As in the 
latter form, it communicates with the exterior by a pair 
of collar-pores which open at the level of the gill-slits. 


The collar-coelom is continued posteriorly into the opercu- 
lum, and anteriorly into the twelve tentacular appendages. 

Finally, behind the collar comes the region of the body 
containing the viscera, which are surrounded by the third 
section of the coelom. 

Only the female reproductive organs have been observed 
up to the present time in Cephalodiscus. They occur as 
a pair of gonadic sacs, opening to the exterior on each 
side of the dorsal middle line between the anus and the 
central nervous system. The latter is very simple, being 
represented merely by a dorsal thickening of the ectoderm, 
with nerve-fibres in the region of the collar and posterior 
portion of proboscis. 

Finally, a short notochordal diverticulum projects into 
the base of the buccal shield as in Balanoglossus. 

Rhabdopleura differs considerably from Cephalodiscus 
in many respects, but, nevertheless, has some fundamen- 
tal characteristics in common with it. In Rhabdopleura 
the individuals of a colony are not independent, but are 
connected with each other by a common cord or caulns, 
which represents the remains of the contractile stalks of 
the polyps. As the growth of the colony proceeds, the 
distal portions of the stalks (i.e. the portions farthest away 
from the animals) become shrunken and hard. The buds 
arise from the soft portions of the caulus, and never be- 
come detached as they do in the case of Cephalodiscus. 
There is only a single pair of tentacular plumes in Rhab- 

FOWLER has recently shown that in Rhabdopleura the 
coelom, whose existence was first established by LAN- 
KESTER, exhibits the same subdivisions as have been 
mentioned above for Cephalodiscus; namely, (i) the cavity 
of the large buccal shield, (2) the collar-cavity opening 


to the exterior by a pair of dorsally placed collar-pores, 
and (3) the body-cavity proper surrounding the alimentary 
canal. According to Fowler, who has recently described 
them in Rhabdopleura, the nervous system and notochord 
have essentially similar relations to those which obtain in 
Cephalodiscus, but there are no proboscis-pores and no 


In the previous pages a good deal of stress has been 
laid on the existence of a praeoral lobe in the various types 
considered. We have recognised it in the snout of Am- 
phioxus (praeoral coelom + praeoral pit), in the proboscis 
of Balanoglossus, the fixing organ of the Ascidian tadpole, 
and in the buccal shield of Cephalodiscus and Rhabdo- 

From a morphological standpoint the praeoral lobe is 
probably one of the most important, as it is certainly one 
of the oldest, structures of the body of bilateral animals, 
and it becomes, therefore, a matter of the first moment to 
be able to trace the modifications which it has undergone 
along the different lines of evolution which have culmi- 
nated in the existing types of animal life. The subject is 
a very large one, and can only be treated here in its 
broadest outlines. 

It is now very generally admitted by zoologists that the 
Echinoderms (star-fishes, sea-urchins, etc.) owe the radial 
symmetry, which is one of the most obvious characteristics 
of their organisation, to their having been derived from 
bilaterally symmetrical ancestors, which became adapted 
to a fixed or sessile existence. If this view is correct, 
and there is good reason for supposing that it is, it follows 
that the majority of living Echinoderms have secondarily 


lost their sessile mode of existence, and have again become 
free-living, retaining, however, their radial symmetry. At 
the present time the fixed habit of life is only retained 
by the members of one of the subdivisions of the Echino- 
derm class ; namely, the Crinoidca. 

Most genera of Crinoids (RJiizocrinus, Pentacrinus, etc.) 
remain fixed by a long, jointed stalk throughout life ; but 
the well-known "feather-star," Antedon rosacea, is only 
fixed during a certain period of its larval development. At 
the close of the period of fixation the body of the animal, 
or, as it is called, the calyx, breaks away from the stalk by 
which it was attached to the rocks, and so begins to lead a 
free existence, being capable of swimming vigorously by 
the flapping of its arms. 

Although the existing Crinoids have become extensively 
modified along their particular line of evolution, yet there 
is reason to believe that they represent the more im- 
mediate descendants of the primaeval form which ex- 
changed its primitively free life and bilateral symmetry for 
a sessile existence and radial symmetry. This view is 
strengthened by the character of the free-swimming larva 
of Antedon. This larva does not possess, in any extrava- 
gant degree, those fantastic structures which are so 
characteristic of other Echinoderm larvae, such as the 
provisional ciliated processes or arms of the " Pluteus " 
(larva of sea-urchins), or the undulating ciliated bands of 

On the contrary, the larva of Antedon is a simple 
barrel-shaped organism, with regular ciliated bands pass- 
ing around it (Fig. 128). 

Perhaps the structure which, above all, stamps the free- 
swimming larva of Antedon as having, from a phylogenetic 
point of view, a more primitive type of organisation than 



that of other Echinoderm larvae, is the well-developed 
apical plate at its anterior extremity. We may express 
this in other words by saying that the larva of Antedon 
possesses a central nervous system at the apex of its 
praeoral lobe. That the prae- 
oral lobe in this larva is not 
sharply marked off from the 
rest of the body is a detail 
of no morphological signifi- 

The apical nervous sys- 
tem of the Antedon larva 
was discovered in 1888 by 

H. BURY, and has been Pig. 128. Free-swimming larva of 
11 i i , Antedon rosacea. from the ventral side. 

more clearly brought out (After SEELIGER .) 

and emphasised in a recent <*/ A P' cal P le - c - b - Ciliated bands. 

f. Fixing disc. v. Vestibulum (so-called 
Work by Dr. OSWALD SEELI- larval mouth, although at this stage 

GER. At the point which is simply an ectodermic g roove )- 
marked externally by the anterior tuft of long cilia in 
Fig. 129 there is a slight groove in the ectoderm below 
which nerve-fibres and ganglion-cells can be identified. 
Seeliger further describes a pair of longitudinal nerves 
running from the nervous area of the apex along the 
ventro-lateral margins of the body. 

As already indicated, the apical plate is, as a general 
rule, conspicuous by its absence in the typical Echinoderm 
larva. In the free-swimming larva of Antedon, however, 
it is emphatically present, although destined to become 
entirely aborted after the fixation of the larva. 

In most Invertebrate larvae in which an apical plate is 
present (e.g. the Trochophore-larva of Annelids and Mol- 
luscs) it becomes, during the metamorphosis, involved in 
other ectodermic thickenings of the prasoral lobe, which 


collectively give rise to the cerebral or supraoesophageal 
ganglion. The apical plate may thus be defined as a 
primitive central nervous system at the apex of the 
praeoral lobe, being the forerunner and formative centre 
of the cerebral ganglion of the Invertebrates. 

Although, with the exception of the Crinoids, there is 
no apical plate in the typical Echinoderm larva, yet, as 
noted above, in many cases a curious transitory lengthen- 
ing of the ectodermic cells at the apical pole has been, 

and can be without great 
difficulty, observed in larvae 
of star-fishes and sea-urchins. 
This alone would seem to 
indicate the former exist- 
ence of a central nervous 
system at the apex of the 
prasoral lobe in the bilateral 

'*- \ x ^y/ / / ancestor of the Echinoderms. 

The way in which the 
Fig. 129. Larva of Asterina gibbosa, primary blastocoelic cavity 

viewed as a transparent object from the 

leftside. (After LUDWIG.) of the prseoral lobe can be 

ent.c. Enteric cavity U Left entero- replaced b dilatation of 

coel, communicating with the right entero- J 

ccei through /./, the prasoral lobe. si. the enterocoel has been de- 


scribed above, both for Tor- 

naria and for the larva of Astcrias vnlgaris (Figs. 121-122). 
In some cases, as in Asterina gibbosa, the praeoral lobe is 
occupied by the enterocoel from the very beginning. In the 
"Pluteus" larva of the Echinids (sea-urchins) the praeoral 
lobe is much reduced ; but in other Echinoderms, as in 
the singular larva of Asterina gibbosa, and in the so-called 
Brachiolaria-larva of the Asterids (star-fishes) in general, it 
is very prominent, and serves as an effective locomoton 
(creeping] organ. 


The very interesting observation has recently been 
made by MACBRIDE, that the larva of Astcrina gibbosa 
actually undergoes temporary fixation at the beginning of 
the metamorphosis, the fixation being effected by the 
praeoral lobe in a manner strikingly similar to that of the 
larvas of Ante don and of dona. 

In the larva of Antedon the adhering disc, by which the 
larva eventually fixes itself to some foreign surface, is 
placed near the front end of 
the prseoral lobe immediately 
below the apical plate. 

The central nervous sys- 
tem of the adult Echinoderm 
arises in entire indepen- 
dence of the actual or sup- 
pressed apical nervous sys- 

Fig. 130. Larva of Asterina gibbosa, 

tem Of the larva, and not at viewed as an opaque object from the left 
,, f ., j r . i side. (After LUDWIG.) 

all from the ectoderm of the j r j PnEoral lobe . 
praeoral lobe. 

We have thus seen how within the limits of a single 
group (viz. the Echinoderms) the praeoral lobe can become 
completely emancipated from the central nervous system ; 
and we have further recognised the fact that whether the 
cavity of the praeoral lobe is a derivative of the primary or 
secondary body-cavity, whether it contains loose mesen- 
chyme or is lined by an endothelium, the morphological 
value of the praeoral lobe itself remains the same. 


It is probable that the misunderstandings and disagree- 
ments which are of such frequent occurrence among mor- 
phologists with regard to the comparison of the types of 
central nervous system presented respectively by the 


Vertebrates and the Invertebrates, are largely clue to the 
failure to detect some general principle of evolution to 
which that archaic structure, the praeoral lobe, has been 

Nevertheless, there are many indications which point 
irresistibly to the conclusion, which I have recently 
brought forward, that the prime factor which must be 
recognised in the evolution of the prasoral lobe, from the 
relations which it presents in the Invertebrates to those 
which it holds in the Protochordates and Vertebrates, is 
its emancipation from the central nervous system. 

In the great groups of the Annelids, Molluscs, and 
Arthropods, the praeoral lobe (prostomium, procephalic 
lobe) is essentially the seat of the brain or cerebral gan- 
glion. The latter, through its representative, the apical 
plate, is the main and often the sole element of the central 
nervous system in the Trochophore-larva of Annelids and 

* In speaking of the apical plate as the forerunner or formative centre of 
the cerebral ganglion, it must not be assumed that these are not distinct 
structures. The apical plate is essentially median and unpaired, while the 
cerebral ganglion is paired. They can both, however, be included under the 
general term, apical nervous system, since they arise from the ectoderm of 
the praeoral lobe. On the other hand, the cerebral ganglion may arise inde- 
pendently of an apical plate; as, for instance, in Lumbricus, where there is 
no apical plate, or in the Nemertines, where the apical plate is discarded 
together with other larval structures (Pilidium). Again, as in Lumbricus and 
many other cases, the cerebral ganglion, after having separated from the 
ectoderm of the praeoral lobe, may recede backwards for a considerable dis- 
tance, so as not to lie in the pneoral lobe in the adult. It is possible that the 
position of the cerebral ganglia of Nemertines may be accounted for by some 
such phylogenetic recession from the praeoral lobe. 

If necessary, it might be said that the prreoral lobe can acquire emancipa- 
tion from the central nervous system by a simple recession of the cerebral 
ganglion. In the case of the Protochordates, however, on the view here advo- 
cated, the praeoral lobe has acquired emancipation from the central nervous 
system, not by the mere recession, but by the complete disappearance of the 
Invertebrate cerebral ganglion. 


At a later stage of development the longitudinal nerve- 
cord (confining the description to the Annelids for the 
sake of simplicity) arises independently of the cerebral 
ganglion, from a pair of longitudinal thickenings of the 
ectoderm near the mid-ventral line, becoming secondarily 
connected with the cerebral ganglion by the circumoesoph- 
ageal nerve-collar or commissure. 

As already indicated, it seems probable, as was sug- 
gested by BALFOUR and GEGENBAUR, that the ventral 
nerve-cord of the Annelids is to be regarded as having 
arisen phylogenetically by the mutual approximation of 
two such lateral cords as occur in the Nemertines, and 
like the latter may be supposed to have originated by a 
concentration on the ventral side of the body of that 
primitively continuous sub-epidermic nerve-plexus which 
is such a characteristic feature of the Nemertines. From 
a consideration of the adult nervous system in the 
Echinoderms, Nemertines, Enteropneusta (Balanoglossus), 
Annelids, and Molluscs, it is evident that such a con- 
centration of nervous tissue has from first to last occurred 
along very different lines. 

Speaking in broad terms, it may be said that the only 
portion of the Invertebrate nervous system which, in its 
prime essence, is invariable and universal (due allowance 
being made for exceptional cases) is the cerebral ganglion 
or its forerunner, the apical plate, the seat of which lies in 
the praeoral lobe. 2 

Under these circumstances it will suffice to confine our 
attention to the prasoral lobe, in the belief that if an 
understanding can be arrived at with regard to that impor- 
tant structure, one of the chief difficulties in the way of a 
just conception of the relations existing between Verte- 
brates and Invertebrates will have been overcome. 


Returning now to Balanoglossus, we have to remark 
that in the Tornaria larva the central nervous system is 
represented entirely by the apical plate of the praeoral 
lobe, the situation of the apical plate corresponding to the 
anterior tip of the proboscis of the adult. Unlike the 
Annelids, however, the apical plate of Tornaria does not 
become replaced after the manner of the Invertebrates by 
the development of a cerebral ganglion arising like it from 
the ectoderm of the praeoral lobe and with it as a formative 
centre. On the contrary, it completely disappears after 
the metamorphosis, having become replaced physiologically 
by the development of the medullary tube in true Verte- 
brate fashion from the dorsal ectoderm of the collar-region 
behind the praeoral lobe.* 

In the Ascidian larva, however, and in Amphioxus, the 
characteristic Invertebrate apical nervous system no longer 
appears in any stage of development, its physiological func- 
tion having been once for all assumed by the medullary 
tube (cerebral vesicle + spinal cord) which lies par excel- 
lence behind the praeoral lobe (Fig. 131). 

Anterior and Posterior Nenrenteric Canals, and the 
Position of the Mouth in the Protochordates. 

After the postoral medullary tube had led indirectly to 
the complete obliteration of the praeoral apical nervous 
system, and had attained to such a degree of development 
as we find, for instance, in the Ascidian tadpole, the central 
canal of the cerebro-spinal nervous system appears to 
have acquired remarkable relations with the alimentary 
canal. At both ends of the body connecting ducts be- 

* For a detailed account of the formation of the medullary tube in the col- 
lar-region of Balanoglossus see MORGAN (Bibliography, Nos. 124 and 125). 



came established between the nervous and digestive 
systems, known respectively as the anterior and posterior 
neurenteric canals. 

The posterior neurenteric canal is only of transitory 
occurrence in all existing Vertebrates, and leads from the 

Fig. 131. Diagrammatic representations of the anterior region of the body 
in (A) an Ascidian larva, (B) larva of Amphioxus, and (C) Balanoglossus. 
(After WiLLEY.) 

The figure of Balanoglossus was compiled from Bateson's figures; the pro- 
boscis-pore is indicated rather too far forwards. 

p.l. Praeoral lobe (fixing organ, snout, proboscis), end. Endostyle. p.p. Praeoral 
pit or proboscis-pore, m. Mouth, np. Neuropore. nc. Medullary tube. ch. Noto- 
chord. e. Eye. of. Otocyst. gl. and h. Proboscis-gland and proboscis-heart of 




-0. < 

Fig. 132. Sagitta hexaptera from the 
ventral surface ; nearly three times natural 
size. (After O. HERTWIG.) 

a. Anus. be 1 . Head-cavities. be 1 -. 
Trunk-coelom. 6c 3 . Caudal coelom. c.l. 
Caudal septum, corn. Commissure, from 
the cerebral ganglion to the single ventral 
ganglion. f l ,f 2 ,f z - Fins. m. Mouth. 
o.d. Oviduct. ov. Ovary, sp. Prehen- 
sile bristles, s.v. Seminal vesicle, t. Tes- 
tis. v.g. Ventral ganglion. 

neural tube into the extreme 
posterior end of the aliment- 
ary canal ; in fact, into that 
portion of it which, in the 
embryos of the higher forms, 
is known as the post-anal 
gut. The anterior neuren- 
teric canal, in its most primi- 
tive condition, opens into the 
base of the buccal tube 

(Fig. 131). 

On this account we find 
in the Ascidian tadpole that 
the mouth is no longer ven- 
tral, as it is in Balanoglossus, 
but is placed dorsally, im- 
mediately in front of the 
anterior extremity of the 
medullary tube. This in- 
timate relation between the 
mouth and the central ner- 
vous system gives a reason 
for the contrast between the 
dorsal position of the mouth 
in the Ascidian tadpole and 
its ventral position in Bala- 

In Amphioxus we have 
seen that the mouth has been 
forced aside from its more 
primitive dorsal position by 
the forward extension of the 
notochord to the tip of the 


praeoral lobe. The origin of the main cavity of the prae- 
oral lobe in Amphioxus from the right of a symmetrical 
pair of head-cavities (anterior intestinal diverticula of 
Hatschek) has been described in a previous chapter. In 
Balanoglossus there is no such complete division of the 
praeoral body-cavity, but it is throughout a single space, 
its right and left halves being confluent. If we now com- 
pare the condition of things in the embryo of Amphioxus, 
where we have a symmetrical pair of head-cavities, with 
that of some other form which, in the adult condition, 
possesses a distinct pair of such cavities, it may assist us 
in imagining how the mouth could have assumed such 
opposite relations as have been mentioned above. 

But first it may be pointed out that in Appendicularia, 
where, as it would appear, in correlation with the second- 
ary acquirement of a purely pelagic habit of life (although 
this point of view is not shared by such authorities as 
Herdman, Seeliger, and Brooks), the praeoral lobe has 
been reduced to a minimum, or to zero, the mouth has 
thereby come to lie in a terminal, or sub-terminal, position, 
with a slight tendency towards the dorsal side.* 

In the curious pelagic worm, Sagitfa, we meet with 
another instance of an animal in which the praeoral lobe, 
in the ordinary sense of the term, is reduced to a mini- 
mum, and the mouth has therefore a sub-terminal position, 
with a ventral inclination (Fig. 132). But although there 
is no distinct praeoral lobe in Sagitta, there is, neverthe- 
less, a pair of head-cavities, which are directly comparable, 
if not perfectly homologous, with the above-mentioned 

* Whatever the truth may be as to the precise systematic position and 
phylogenetic value of Appendicularia, one thing, to my mind, remains abso- 
lutely certain, namely, that it has descended from a form which possessed a 
praeoral lobe, and that it has secondarily lost that structure. 


head-cavities of Amphioxus, although they have a some- 
what different origin. 

It should not be forgotten that Sagitta occupies a very 
isolated position in the zoological system, being placed in 
a group by itself, the Chatognatha, and that therefore the 
peculiarities of its organisation cannot be taken as repre- 
senting any definite intermediate stage in the phylogeny 
of other forms, yet, from a general standpoint, the con- 
ditions which it presents in its life-history are highly 

The head-cavities of Sagitta arise by constriction from 
the anterior extremities of the single pair of archenteric 
pouches which give rise to the coelom of the adult. They 
remain distinct and separate on either side of the head 
throughout life. If, now, we imagine them to grow for- 
ward and fuse together in front of the mouth, in a simi- 
lar manner to that described above for the enteroccelic 
pouches of Asterias, we should have a praeoral body-cavity 
of a similar character to that of Balanoglossus. 

Now, the ultimate position of the mouth under these 
new conditions would depend upon circumstances affect- 
ing the whole organisation of .the animal. 

In an animal whose grade of organisation was on an 
approximate level with that of Sagitta the mouth would 
undoubtedly remain on the ventral side of the body. But 
in an animal whose organisation had reached the stage 
of evolution represented by that unknown ancestor of 
Amphioxus (most nearly represented at the present time 
by the Ascidian tadpole), whose notochord did not extend 
beyond the anterior limit of the neural tube, the mouth 
would pass to the dorsal side of the body to come into 
connexion with the neural canal. 



After what has been said above, in this and the preced- 
ing chapters, the question as to how the praeoral lobe is 
represented in the craniate Vertebrates need not detain us 

Since, as shown above, the nervous element of the prae- 
oral lobe (apical plate and cerebral ganglion) is entirely 
lacking in the Vertebrates, we can only expect to find the 
mesodermal element represented in the head-cavities of 
the higher forms. 

In consequence of the great development of the brain, 
even in the lowest craniate Vertebrates, as compared with 
Amphioxus, and in consequence too of the cranial flexure, 
the head-cavities have been made to assume a more sub- 
ordinate position, and no longer take part in the formation 
of a prominent lobe in front of the body. This is a perfect 
illustration of "le principe du balancement des organes " 
of Geoffroy Saint-Hilaire, the praeoral lobe decreasing as 
the brain increases. A comparison between Figs. 70, 
72, 117, and 135 will show at once that the praeoral head- 
cavities of Amphioxus and Balanoglossus are the homo- 
logues of the pr&mandibular head-cavities of the craniate 

These cavities lie at first below the mid-brain, and later 
their walls give rise to most of the eye-muscles. In Figs. 
91 and 135 the median portion of the praemandibular 
cavities can be seen still in the form of an anterior pocket 
of the endoderm, and it may be noticed how far it is 
removed from the anterior extremity of the body to which 
it extends in Amphioxus, etc. In the craniate Verte- 
brates the brain extends forwards, and the head-cavities 



remain behind. This is, as we should expect, the exact 
reverse to what obtains in Amphioxus. 

In connexion with the evolution of the praeoral lobe, 
we thus have an excellent example of repeated change of 

We may conclude, therefore, that the prasoral lobe, 
which, in the Invertebrates, is above all the bearer of the 
cerebral ganglion, and in the Protochor dates is released 
from this function and becomes in part a locomotor 
(Balanoglossus, Cephalodiscus) fixing (Ascidian) and bur- 
rowing (Amphioxus) organ, is represented in the craniate 
Vertebrates by the prcemandibular head-cavities, whose 
walls give rise to most of the eye-muscles. 


In consequence of the increase in the size of the brain, 
its forward extension and its cranial flexure, together with 
the relative reduction of the head-cavities, it is obvious 
that the mouth has been carried round from its primitively 
dorsal position to its final position on the ventral side of 
the head in the craniate Vertebrates. (Cf. Fig. 91.) This 
would have been all that need be said about the mouth 
were it not for the fact that the view, originally started by 
DOHRN, that the Vertebrate mouth was a new formation 
resulting from the fusion of two gill-slits, has received such 
wide support and still in a measure holds its own. 

Since the Annelid mouth perforates the central nervous 
system in passing through the circumcesophageal nerve- 
collar, it was necessary to frame a theory which would 
get over the difficulty that nothing of the kind occurs in 
the Vertebrates. Accordingly Dohrn supposed that the 
old Annelid mouth had become aborted, and was replaced 



by a new mouth derived from a fusion across the mid- 
ventral line of a pair of gill-clefts. DOHRN was a trifle 
uncertain as to the rudiment of the old mouth, but BEARD 
was more certain on this point, and thought he had estab- 
lished the fact that the hy- 
pophysis cerebri represented 
the remains of the old An- 
nelid mouth. 

Dohrn certainly succeeded 
in bringing forward some 
apparently good evidence in 
support of his theory of the 
gill-slit origin of the mouth. 
This evidence was derived 
from the study of the de- 
velopment of the mouth in 

Teleostean or bony fishes. Fig . I33 ._ TWO frontal views of an 

In manv Teleosteans the ernbrvo of Batrachus tan, to show the 

double nature of the stomodoeum. (From 

mouth has at first an appar- hitherto unpublished drawings kindly lent 
, , , , , ... by Miss C. M. CLAPP.) 

ently double origin, in that The embrvo is lying upon the yolk 
two separate ectodermal in- and the se P tum which divides the stomo- 

doeum passes from the upper lip to the 

growths OCCUr which fuse surface of the blastoderm which covers 

, i , i j j i the yolk. The lower figure is a drawing 

With the endoderm, instead O f th e sa me embryo as the -upper, a few 

of the median stomodoeal hours later - Above the stomodceum are 

seen the small nasal pits (rudiments of 

involution which is SO char- the external nares), and at the sides of 
, r .1 -IT the head are the rudiments of the eves. 

actenstic of other Verte- 
brates. This double origin of the mouth is particularly 
well shown in the embryos of the remarkable toad-fish, 
Batrachus tau, as observed by Miss CORNELIA CLAPP at 
the Marine Biological Laboratory of Woods Roll, Mass., 
in 1 889 (Fig. 133). In this case the mouth-cavity is seen 
to be divided into two halves by a median septum. 

Subsequently the septum becomes absorbed, and the 


two halves of the mouth coalesce. In view of the pre- 
vious existence of the gill-slit theory of the mouth, some 
such theory being a necessary accessory to the Annelid- 
theory, it is not surprising that this undoubted double 
origin of the mouth in Teleosteans should be regarded as 
a striking confirmation of Dohrn's hypothesis. And yet, 
occurring as it does only in the Teleosteans, whose devel- 
opment is admittedly in many respects highly modified, 
the interpretation which Dohrn and his followers have 
placed upon this observation must always have been open 
to doubt. The simplest explanation of the double origin 
of the Teleostean mouth is that, owing to certain condi- 
tions (possibly mechanical) of development, the two angles 
of the mouth develop before the median portion. This is 
the conclusion which H. B. POLLARD has also reached in 
his recent studies on the development of the head in the 
Teleostean fish, Gobins capita. 

According to the standpoint I have adopted in the fore- 
going pages, there is no a priori reason for doubting that 
the Vertebrate mouth is completely homologous with the 
Protochordate mouth ; and that the latter in its turn is 
the direct descendant of the typical Invertebrate mouth. 

Again, the anatomy and development of the Protochor- 
dates and of the Cyclostomi (Ammoccetes) show no indica- 
tion whatever of a discontinuity in the evolution of the 
most highly elaborated mouth of the gnathostomous or 
jawed Vertebrates. 

We conclude, therefore, that the ventral mouth of the 
craniate Vertebrates is the homologue of the primordial 
dorsal mouth as we find it in the Protochordates, and that 
its direction of evolution has been, as was so ably main- 
tained by BALFOUR, from the cyclostomous to the gnatho- 
stomous condition. 



The pituitary body, or hypophysis, belongs to the series 
of ductless "glands" (pineal body, thyroid gland, thy- 
mus, etc.) which are such a characteristic feature of the 
vertebrate organisation. It arises as an ectodermal invo- 
lution from the roof of the stomodceum, directed towards 
the base of the primary fore-brain, from which the infun- 
dibulum grows out. 

The pituitary involution becomes in most forms nipped 
off from the stomodceum, and then lies as a closed sac 
in contiguity with the infundibulum. Later on it produces 
a system of branches, the lumina of which tend to dis- 
appear ; and in some forms (e.g. Mammalia) it undergoes 
actual fusion with the infundibulum. 

The very constant relation of the hypophysis to the 
infundibulum in the craniate Vertebrates (see Fig. 134) 
naturally led to the supposition that there must originally 
have been a functional connexion between the two struct- 
ures of a similar nature to that which exists between the 
olfactory pit and neuropore in Amphioxus. Recent re- 
searches, however, have rendered it probable that such a 
supposition is erroneous. VON KUPFFER has discovered 
the homologue of the lobus olfactorius of Amphioxus in 
the craniate Vertebrates, and has shown that it occurs at 
a point far removed from the infundibular region. 

Until recently it was also very generally thought that 
the infundibulum represented the anterior end of the 
brain, which had become bent downwards and backwards 
by the cranial flexure. Kupffer, however, has brought for- 
ward weighty reasons for doubting this view. According 
to him, the infundibulum is essentially a downgrowth or 



evagination from the floor of the brain, occurring behind 
the anterior terminal extremity of the brain. 

It follows that the morphological anterior extremity of 
the craniate brain coincides with the median lobus olfac- 
torins inipar, which also represents the point of last con- 
nexion of the medullary tube with the superjacent ecto- 
derm. The lobus olfactorius impar lies in the anterior 
vertical wall, which forms the boundary of the primary 
fore-brain in front, known as the lamina tcrminalis. RABL- 
RUCK.HARD has also observed the median olfactory lobe in 

Fig. 134. Sagittal section through the head of an embryo of Acanthias.. 

a.c. Position of anterior commissure, al. Alimentary canal, cer. Cerebellum. 
ch. Notochord ; the black shading below the notochord indicates the aorta. 
f.b. Fore-brain, h.b. Hind-brain. Ay. Hypophysis, already shut off from the 
stomodoeum and lying as a closed sac at the base of inf, the infundibulum. 
l.o. Lobus olfactorius. m. Mouth, m.b. Mid-brain, o.c. Optic chiasma. p.b. 
Pineal body (epiphysis). 

the Selachian embryo (Fig. 134), and it has since been 
found by BURCKHARDT in other forms. 

It can thus hardly be doubted that the median rudi- 
mentary olfactory lobe of the embryos of the higher 
Vertebrates is homologous with the lobus olfactorius of 
Amphioxus (Fig. 51), and, like the latter, represents the 
remains of the neuropore. In Amphioxus, however, the 


olfactory lobe abuts against the olfactory pit, and, in fact, 
in young individuals opens into it by the neuropore 

(Fig. 45). 

On the view which I have urged above, that the 
olfactory pit of Amphioxus is homologous with the 
hypophysis cerebri of the craniate Vertebrates, it must 
be assumed that in the latter forms, the neuropore hav- 
ing ceased to be in any way a functional organ, the hy- 
pophysis, which has likewise become (morphologically) a 
vestigial structure, has been mechanically separated from 
the neuropore, with which it was primitively in functional 
connexion. It must be supposed that this separation of 
the hypophysis from the neuropore has been effected by 
the more rapid downward growth of the ectoderm (from 
which the hypophysis arises) than of the wall of the brain, 
so that the hypophysis has been carried farther round to 
the lower side of the head than the neuropore (Fig. 135). 
The reason for this unequal growth of the external body- 
wall and of the cerebral wall may, perhaps, be sought for 
in the great and independent increase in the cubical con- 
tents of the brain. 3 

We thus arrive at the conclusion that the present 
relation of the hypophysis to the infundibulum in the 
craniates, however intimate it may be in some cases, is, 
nevertheless, incidental and secondary. 

That this conclusion is not so strained as might appear 
at first sight is clearly shown by the fact that the in- 
fundibulum is not the only structure with which the 
hypophysis enters into close relations. 

In the exceptional cases of Myxine and Bdellostoma, 
for instance, the distal end of the hypophysis has nothing 
to do with the infundibulum, but actually opens into the 
pharynx. In these hag-fishes, as also in the lamprey 



(where there is no internal opening of the hypophysis 
into the pharynx), the external opening of the hypophysis 
does not close up, as in the higher forms, but persists 
throughout life, becoming carried round to the top of 
the head during the embryonic development by differ- 
ential growth of neighbouring parts, as has been actually 
observed in Petromyzon. 



Fig. 135. Median sagittal section through the head of young Ammocostes. 
(After KUPFFER.) 

The arrow indicates the extent to which the hypophysis has been (hypothetically) 
removed from the neighbourhood of the neuropore (lobus olfactorius impar). 

ch. Notochord. ec. Ectoderm, en. Endoderm. ep. Epiphysis. hy. Hypo- 
physial involution, l.o. Lobus olfactorius impar. n. Nasal involution, pm. Me- 
dian portion of praemandibular cavity, st. Stomodoeum. F.M.H. Primary fore-, 
mid-, and hind-brain. 

In other cases, as, for example, in the embryo of the 
rabbit, it has been observed that the hypophysis actually 
undergoes a temporary fusion with the front end of the 
notochord ; and in all cases the distal end of the hypophysis 
grows inwards as much towards the notochord as towards 
the infundibulum, so that for the embryonic stages of the 
craniate Vertebrates it might be said that the relations of 


the hypophysis to the front end of the notochord are as con- 
stant as its relations to the infundibulum. So close is the 
apparent relation of the hypophysis to the notochord that 
at least one zoologist, HUBRECHT, has suggested that there 
was originally a functional connexion between the two 

Again, in the embryo of Acipenser, the sturgeon, as 
shown by KUPFFER, the distal end of the hypophysis 
undergoes temporary fusion with the subjacent wall of 
the alimentary cavity. In spite of the extremely modified 
character of the embryo of Acipenser (the embryo being 
flattened out like a disc over the yolk), Kupffer regards 
this fusion of the hypophysis with the endoderm as being 
of great morphological significance. 

On the contrary, for the reasons mentioned above, I 
would regard all these fusions of the hypophysis in the 
craniate Vertebrates, whether with the infundibulum, 
notochord, or endoderm, as being of an entirely incidental 
character, often due, perhaps, to a tendency of such con- 
tiguous embryonic tissues to fuse together. 

I therefore suggest that : The liypopJiysis arose in con- 
nexion zvit/i a functional neuropore ; wJien the neuropore 
ceased to be functional, tliere ^vas no longer any bond of 
union between its inner portion, whicJi opened into the 
cerebral cavity, and its outer portion, wJiich opened into the 
buccal cavity ; and these two portions became separated by 
differential growth of the cerebral and body-walls (cf. Fig. 

The Ascidian Hypophysis. 

The development of the hypophysis in a typical As- 
cidian, its constriction from the wall of the cerebral 
vesicle in the form of a tube, and its opening into the 


buccal cavity, or branchial sac, have been described above. 
The most serious objection which has been raised against 
the comparison of the hypophysis of the Ascidians with 
that of the craniate Vertebrates is, that in the former 
the hypophysis opens, not at an ectodermal surface into 
the stomodoeum, but at an endodermal surface (behind the 
stomodoeum) into the branchial sac. This is undoubtedly 
the case in some Ascidians, e.g. Distaplia, and probably 
also in Clavelina, etc. In Ciona, however, as I can state 
after renewed study of the question, it apparently opens at 
first into the buccal cavity precisely in the line of junction 
between the stomodoeum and the branchial sac, so that its 
upper margin is continuous with the stomodoeal epithelium, 
while its lower margin is continuous with the epithelium 
of the branchial sac. 

It is probable that too much stress has been laid on the 
question whether the hypophysis of the Ascidians opens 
at an endodermic or at an ectodermic surface, and that 
thus the attention has been diverted from the essential 
fact that the hypophysis opens into the buccal tube at the 
entrance to the branchial sac. In the case of the Ascid- 
ians, therefore, I should also regard the fusion of the 
hypophysis, whether with the ectoderm of the stomodoeum 
or with the endoderm of the branchial sac, as being in 
itself non-essential, while the actual opening of the hy- 
pophysis (itself derived by constriction from the nerve- 
tube) into the buccal cavity, apart from the question of an 
ectodermal or endodermal surface, is the essential point. 



From the facts that have been recorded and the consid- 
erations that have been urged in these pages, it would 
follow that one of the chief factors in the evolution of the 
Vertebrates has been the concentration of the central 
nervous system along the dorsal side of the body (in 
contrast to the position of the longitudinal nerve-cord of 
Annelids, etc., along the ventral or locomotor surface), and 
its conversion into a hollow tube. If it be admitted that 
the hypophysis became evolved in connexion with a func- 
tional neuropore, it is obviously a structure which has 
arisen within the limits of the Vertebrate phylum, and can, 
therefore, have no representative in the typical Invertebrate 
organisation. It has been suggested by ADAM SEDGWICK 
and VAN WIJHE that the original function of the central 
canal of the spinal cord was to promote the respira- 
tion (oxygenation) of the tissue of the central nervous 
system, water entering by the neuropore, and passing out 
through the posterior neurenteric canal. 

It is not so easy to form a conception as to the prime 
origin of the other two cardinal characteristics of a 
Vertebrate (Chordate) ; namely, gill-slits and notochord. 

As to the origin of gill-slits, it has been suggested inde- 
pendently by HARMER and BROOKS, that they arose at first 
not so much to perform the direct function of respiration, 
as to carry away the bulk of the water which constantly 
entered the mouth with the food, so as to avoid the neces- 
sity and discomfort of the never-ceasing flow of water 
through the entire length of the alimentary canal. In 
Cephalodiscus, for example, the luxuriant branchial plumes 
must be sufficient for the respiration of the minute animal, 


while the usefulness of the pair of gill-slits, in allowing the 
surplus water to pass out of the pharynx, is evident. 

The notochord is more difficult to explain, and the fact 
of its occurrence in the proboscis of Balanoglossus and 
in the tail of the Ascidian tadpole is very puzzling. The 
mode of its occurrence in Balanoglossus is undoubtedly 
divergent, and not in the direct line of Vertebrate descent. 
It is possible that the notochord has not arisen through a 
process of elaborate change of function from a pre-existing 
structure, but simply as a solidification of the endoderm 
which was continued into the caudal or post-anal extension 
of the body to form the axial support for a locomotor tail ; 
while the subsequent extension of the notochord into the 
pras-anal region of the body is not difficult to understand. 
The general capacity of the endoderm for producing 
skeletal tissue is already present in some of the Medusae 
and Hydroid polyps whose tentacles are stiffened by a 
solid endodermal axis. 

From a purely morphological point of view it now 
seems as though the prasoral lobe and in a lesser degree, 
perhaps, the hypophysis, would materially assist in furnish- 
ing the key to a correct appreciation of the relationship 
between the craniate Vertebrates, the Protochordates, 
and the Invertebrates. 

As we have indicated above, in the formulation of the 
Annelid-theory 4 no allowance has been made for the prin- 
ciple of parallelism in evolution ; but it is impossible to 
doubt that this is a very potent factor which should always 
be borne in mind in estimating the genetic affinity between 
widely different groups of animals. The closer the super- 
ficial resemblance between an Annelid and a Vertebrate 
(in the possession of somites, segmental organs, etc.) is 
shown to be, the more perfect appears the parallelism 


in their evolution and the more remote their genetic 

For the present we may conclude that the proximate 
ancestor of the Vertebrates was a free-swimming animal 
intermediate in organisation between an Ascidian tadpole 
and Amphioxus, possessing the dorsal mouth, hypophysis, 
and restricted notochord of the former ; and the myo- 
tomes, coelomic epithelium, and straight alimentary canal 
of the latter. The ultimate or primordial ancestor of the 
Vertebrates would, on the contrary, be a worm-like animal 
whose organisation was approximately on a level with 
that of the bilateral ancestors of the Echinoderms. 


i. (p. 246.) For the discussion of the phenomena of meta- 
merism and the enumeration of examples of independent metameric 
repetition of parts, consult the following : LANG, ARNOLD. Der 
Ban von Gunda Segmentata und die Verwandtschaft der Plathel- 
minthen mit Ccelenteraten und Hirudineen. Mitth. Zool. Stat. 
Neapel, Bd. III. 1882. p. iS-j e/ seg. SEDGWICK, ADAM. On 
the Origin of Metameric Segmentation, and Some Other Mor- 
phological Questions. Quarterly Jour. Micro. Sc. XXIV. 1884. 
pp. 43-82. BATESON, WILLIAM. The Ancestry of the Chordata. 
Quarterly Jour. Micro. Sc. XXVI. 1886. pp. 535-571- CALD- 
WELL, H. Blastopore, Mesoderm, and Metameric Segmentation. 
Quarterly Jour. Micro. Sc. XXV. 1885. pp. 15-28. HUBRECHT, 
A. A. W. Report on the Nemertea collected by H. M. S. Challenger, 
1873-76. Chall. Kept. Zool. XIX. 1886. (Also, HUBRECHT. 
The Relation of the Nemertea to the Vertebrata. Quarterly Jour. 
Micro. Sc. XXVII. 1887. pp. 605-644.) VAN BENEDEN, 
EDOUARD. Recherches sur le Dweloppement des Arachnactis. 
Contribution a la Morphologie des Cerianthides. Archives de 
Biologic, XI. 1891. pp. 115-146. Also consult the recent 
great work of BATESON, Materials for the Study of Variation. 
London, 1894. 


2. (p. 273.) On the subject of the praeoral lobe and the api- 
cal nervous system of Invertebrates, see the following: BALFOUR, 
F. M. Comparative Embryology. 1881. Vol. II. Chap. 12. 
Observations on the Ancestral Form of the Chordata. BEARD, 
J. The Old Mouth and the New, A Study in Vertebrate Mor- 
phology. Anat. Anz. III. 1888. pp. 15-24. WILSON, E. B. 

The Embryology of the Earthworm. Jour. Morph. III. 1889. 
pp. 387-462. HATSCHEK, B. Lehrbuch der Zoologie. 3d Liefer- 
ung. Jena, 1891. WILLEY, A. On the Evolution of the Praoral 
Lobe. Anat. Anz. IX. 1894. pp. 329-332. 

3. (p. 285.) From what has been said in the text, it is obvious 
that the hypophysis of the craniate Vertebrates, in becoming 
separated from the neuropore, has retained (at least in the embryo) 
its primitive relations with the buccal cavity, and, like the latter, 
has been made to assume its present position in consequence of 
the forward growth of the brain and the ensuing cranial flexure. 
In Amphioxus, the hypophysis (i.e. olfactory pit) arises as an 
ectodermic involution immediately over the neuropore, but still 
independent of the latter. In other words, the neuropore exists 
in Amphioxus for a considerable length of time before the hypoph- 
ysis forms ; and this is in accordance with what we should expect 
from the analogy of the craniate Vertebrates. In the Ascidians, 
however, the conditions are somewhat different, and there is at first 
no such obvious differentiation between neuropore and hypoph- 
ysis. For the simple Ascidians (e.g. Ciona) it must at present 
remain doubtful whether the increase in size of the hypophysis 
takes place entirely by interstitial growth, or whether there is any 
ingrowth from the wall of the buccal tube at the lips of the aper- 
ture (dorsal tubercle) of the hypophysis. In any case there are 
not wanting indications in the Ascidians of a distinction, and even 
separation, between the distal portion of the hypophysis, which 
at first opens into the cerebral vesicle, and the proximal portion, 
which opens into the buccal cavity. In the adult, the proximal 
portion of the hypophysis has the form of a simple duct, opening 
by the so-called dorsal tubercle into the buccal cavity, while the 
subneural gland arises as a proliferation from the ventral wall of 
the distal portion. In Phallusia mammillata, as was discovered 
by JULIN {Archives de Biologie, II. iSSi. pp. 211-232), num- 

NOTES. 293 

bers of secondary tubules grow out from the principal duct of the 
hypophysis, and acquire ciliated funnel-like openings into the 
peribranchial chamber ; subsequently HERDMAN (Proc. Roy. Soc. 
Edinburgh, XII. 1882-84. p. 145) found that in this form the 
dorsal tubercle, or opening of the hypophysis into the buccal cavity, 
is sometimes absent. In Ciona intestinalis I have found in young 
individuals an obliteration of the lumen of the hypophysis between 
the proximal and the distal portions. In other cases, as in Appen- 
dicularia, the glandular portion of the hypophysis may be reduced 
or absent. 

On the subject of the Ascidian hypophysis, the following papers 
should also be consulted : SHELDON, LILIAN. Note on the Ciliated 
Pit of Ascidians and its Relation to the Nerve-ganglion and So- 
called Hypophysial Gland. Quarterly Jour. Micro. Sc. XXVIII. 
1888. pp. 131-148. HJORT, JOHAN. Ueber den Entwicklungs- 
cyclus der Zusammengesetzten Ascidien. Mitth. Zool. Stat. Neapel, 
X. 1893. pp. 584-617. METCALF, MAYNARD M. The Eyes and 
Subneural Gland of Salpa. Baltimore, 1893. (Published as 
Part IV. of Professor Brooks's Monograph of the Genus Salpa.) 

4. (p. 290.) The most complete presentation of the Annelids- 
theory is contained in the classical Monographic der Capitel- 
liden des Golfes von Neapel, by Dr. HUGO EISIG. It is needless 
to add that this monograph will command the gratitude and 
admiration of zoologists to the end of time. 



1 CARUS, J. VICTOR. Geschichte der Zoologie. Munchen, 1872. 

2 DOHRN, ANTON. Der Ursprung der IVirbelthiere tend das Prin- 
cip des Functionsivechsels. Leipzig, 1875. 

3 HAECK.EL, ERNST. Anthropogenic oder Entiuickelungsgeschichte 
des Menschen. Leipzig, 1874; 4th Edit., 1891. 

4 LANKESTER, E. RAY. Article " Verfebrata." Encycl. Brit., 
9th Edit. Republished in "Zoological Articles," London, 1891. 

5 PERKIER, EDMOND. La Philosophie Zoologique avant Darwin, 
2d Edit. Paris, 1886. 

6 SEMPER, CARL. Die Verivandtschaftsbeziehungen der geglieder- 
ten Thiere. Parts I. to III. Wlirzburg, 1875-76. 


7 ANDREWS, E. A. The Bahama Amphioxns (preliminary ac- 
count). Johns Hopkins University Circulars, Vol. XII. p. 104. 
June, 1893. 

8 ANDREWS, E. A. An Undescribed Acraniate: Asymmetron 
lucayanum. Studies from the Biol. Lab. Johns Hopkins Uni- 
versity, Vol. V. No. 4. 1893. pp. 213-247. Plates XIII.- 

Contains bibliography of systematic and faunistic works on 

9 ANTIPA. GR. Ueber die Beziehungen der Thymus zu den soge- 
nanntcn Kiemenspaltenorganen bei Selachiern. Anat. Anz. 
VII. 1892. pp. 690-692. One figure in text. 

* This bibliography does not by any means include all that has been written 
on the anatomy of Amphioxus. Some of the older and shorter works, as well 
as some of those relating to special points of histological detail, have been omitted, 
as they are fully dealt with in many of the memoirs here cited. 



10 BALFOUR, F. M. A Preliminary Account of the Development of 
the Elasmobranch Fishes. Quarterly Jour. Micro. Sc. XIV. N. S. 

1874. pp. 323-364. Plates 13-15. 

Paper in which Balfour first published his discovery of the seg- 

mental origin of excretory tubules. This was made out also in the 

same year by Semper and Schultz. (Vide infra, Schultz.~) 

n BALFOUR, F. M. On the Origin and History of the Urino- 

genital Organs of Vertebrates. Jour, of Anat. and Physiol. X. 

1875. PP- I 7~4^- Eight figures in text. Amplification of his pre- 
vious work, with bibliography up to date. 

12 BALFOUR, F. M. The Development of Elasmobranch Fishes. 
Development of the Trunk. Jour, of Anat. and Physiol. XI. 

1876. pp. 128-172. Plates 5 and 6. First account of origin of 
paired limbs from continuous epiblastic thickenings. 

13 BALFOUR, F. M. A Monograph on the Development of Elasmo- 
branch Fishes. London, 1878. 

14 BEDDARD, FRANK EVERS. On the Occurrence of Numerous 
Nephridia in the Same Segment in Certain Earthworms, and on 
the Relationship between the Excretory System in the Annelida and 
in the Platy helminths. Quarterly Jour. Micro. Sc. XXVIII. N. S. 
1888. pp. 397-411. Plates 30-31. Contains discovery of neph- 
ridial network in Perichasta. 

15 BENHAM, W. BLAXLAND. The Structure of the Pharyngeal 
Bars of Amphioxus. Quarterly Jour. Micro. Sc. XXXV. N. S. 
1893. pp. 97-118. Plates 6-7. 

1 6 BOURNE, ALFRED GIBBS. Contributions to the Anatomy of 
the Hirudinea. Quarterly Jour. Micro. Sc. XXIV. N. S. 1884. 
pp. 419-506. Plates 24-34. 

Contains discovery of nephridial network in Pontobdella. 

17 BOVERI, THEODOR. Ueber die Niere des Amphioxus. Miin- 
chener Medicin. Wochenschrift. No. 26. 1890. Sep. Abd. 
pp. 1-13. Two figures in text. (Preliminary note.) 

1 8 BOVERI, THEODOR. Die Nierencandlchen des Amphioxus. Ein 
Beitrag zur Phylogenie des Urogenitalsy stems der U'irbelthiere. 
Zoolog. Jahrblicher. Abth. fiir Morphol. V. 1892. pp. 429-510. 
Taf. 31-34 and five figures in text. 

19 COSTA. O. GABRIELE. Cenni zoologici ossia descrizione som- 
maria delle specie nuove di animali discoperti in diverse contrade 
del regno neW anno 1834. Napoli, 1834. See also Fauna del 
regno di Napoli. 1839-50. 

20 CUENOT, L. Etudes sur le sang et les glandes lymphatiques 
dans la serie animale. Archives de zool. expe"rimentale, XIX. 
1891. Amphioxus. pp. 55-56. 


Notes absence of blood-corpuscles in Amphioxus. Those 
described by previous authors must therefore require another ex- 

21 DOHRN, ANTON. Studieti zur UrgescJiichte des ll'irbelthier- 
korpers. IV. Section 5. Entstehung itnd Bedeutung der Thymits 
der Selachier. Mitth. Zool. Stat. Neapel. V. 1884. pp. 141-151. 
Taf. 8. Figs, i and 2. 

22 EisiG, HUGO. Die Segmentalorgane der Capitelliden. Mitth. 
Zool. Stat. Neapel. I. 1879. pp. 93-118. Taf. IV. 

Discovery of numerous nephridia in single segments and an- 
astomoses between successive nephridia. 

23 EMERY, CARLO. Le specie del genere Fierasfer nel Golfo di 
Napoli. 2d Monograph in the " Fauna und Flora des Golfes von 
Neapel." Leipzig, 1880. 

24 EMERY, CARLO. Zur Morphologie der Kopfniere der Teleostier. 
Biologisches Centralblatt, I. 1881. pp. 527-529. See also 
Zoologischer Anzeiger, VIII. 1885. pp. 742-744. 

25 FUSARI, ROMEO. Beitrag sum Studiuin des peripherischeu 
Nervensy stems von Amphioxus lanceolatus. Internationale Mo- 
natsschrift fiir Anatomic und Physiologic, VI . 1889. pp. 120-140. 
Taf. VII.-VIII. 

26 GOODSIR, JOHN. On the Anatomy of Atnphioxus lanceolatus. 
Transactions of the Royal Society of Edinburgh, Vol. XV. Part I. 
1841. pp. 241-263. 

27 GRENACHER, H. Beitrage zur n'dhern Kenntniss der Muscu- 
latur der Cyclostomen und Leptocardier. (Leptocardia proposed 
by Haeckel as a classificatory name on account of the simple 
tubular " heart " of Amphioxus.) Zeitschr. fiir Wiss. Zoologie, 
XVII. 1867. pp. 577-597. Taf. XXXVI. First isolation of 
muscle-plates of Amphioxus. 

28 GUNTHER, ALBERT. Synopsis of Genus Branchiostoma. In 
Report on Zool. Collections of H. M. S. Alert. 1881-82. pp. 31- 
33. London, 1884. 

29 HATSCHEK, BERTHOLD. Die Metamerie des Amphioxus und 
des Ammocastes. Verh. Anat. Gesellschaft, 6th Versammlung. 
Wien, 1892. pp. 137-161. Eleven figures in text. 

29 bis. HATSCHEK, BERTHOLD. Zur Metamerie der Wirbelthiere. 

Anat. Anz. VII. Dec. 1892. pp. 89-91. 

30 HUXLEY, T. H. Preliminary Note upon the Brain and Skiell 
of Amphioxus lanceolatus. Proceedings of the Royal Society, 
XXIII. 1874. pp. 127-132. 

Points out that in Myxine and Ammocoetes a velum is present 
separating the buccal (stomodoeal) from the branchial cavity. 


The resemblance of the buccal cavity and tentacles (cirri) of 
Ammoccetes to the corresponding parts in Amphioxus is so close 
that there can hardly be any doubt the two are homologous. The 
anterior end of the nerve-tube of Amphioxus corresponds to the 
lamina tenninalis of the craniate Vertebrates. 

31 HUXLEY, T. H. On the Classification of the Animal Kingdom. 
Journal of the Linnaean Society (London), XII. 1876. pp. 199- 
226. (Read 3d Dec., 1874.) 

Section on " epiccel? 1 p. 216 et seq. Atrial cavity of Amphi- 
oxus and Ascidians is an epicoel like the opercular cavity of the 
Amphibian tadpole. 

32 KOLLIKER, ALBERT. Ueber das Geruchsorgan von Amphioxus. 
Muller's Archiv fiir Anat. Physiol., etc. 1843. PP- 3 2 ~35- Taf. 
II. Fig. 5. 

Discovery of olfactory pit and first description of the spermatozoa 
of Amphioxus. 

33 KOPPEN, MAX. Beitr'dge ztir vergleichenden Anatomic des 
Centralnervensystems der Wirbelthiere. Zur Anatomic des 
Eidechsengehirns. Morphologische Arbeiten (Schvvalbe), I. 1892. 
pp. 496-515. Taf. 22-24. 

Contains discovery of giant-fibres in caudal portion of spinal 
cord of Lacerta viridis. 

34 KOHL, K. Einige Bemerkungen uber Sinnesorgane des Amphi- 
oxus lanceolatus. Zool. Anz. 1890. pp. 182-185. 

States that sometimes there is a shallow olfactory groove on the 
right side as well as that in the left. Such grooves are often due 
to artificial crumpling, and the observation requires confirmation. 

35 KRUKENBERG, C. FR. W. Zur Kenntnis des chemischen Baues 
von Amphioxus lanceolatus und der Cephalopoden. Zool. Anz. 
iSSi. pp. 64-66. See also HOPPE-SEYLER'S reply, pp. 185-187. 
Compare also CUENOT (supra). 

36 KUPFFER, CARL VON. Studien sur vergleichende Entiuick- 
lungsgeschichte des Kopfes der Kranioten, I. Die Entivicklung des 
Kopfes von Acipenser sturio an Medianschnitten untersucJit. 95 
pp. 8. 9 Tafeln. Miinchen und Leipzig, 1893. 

Contains also a chapter on brain of Amphioxus, with figures. 

37 LANGERHANS, PAUL. Zur Anatomic des Amphioxus lanceolatns. 
Archiv fiir mikroskopische Anatomie, XII. 1876. pp. 290-348. 
Taf. XI I. -XV. 

Standard work on the histology of Amphioxus. 

38 LANKESTER, E. RAY. On Some Neiv Points in the Structure of 
Amphioxus and their Bearing on the Morphology of Vertebrata. 
Quarterly Jour. Micro. Sc. XV. N. S. 1875. pp. 257-267. 


39 LANKESTER, E. RAY. Contributions to the Knowledge of Amphi- 
oxus lanceolatus, Yarrell. Ib., Vol. XXIX. 1889. pp. 365-408. 
Five plates. 

40 LWOFF, BASILIUS. Uber den Zusammenhang von Markrohr 
und Chorda beim Amphioxus und iihnliche Verhdltnisse bei 
Anneliden. Zeitschrift fur wiss. Zoologie. Bd. 65. 1893. pp. 
299-308. Taf. XVII. 

Describes those supporting fibres of the spinal cord of Amphi- 
oxus which descend in successive paired groups to the notochordal 
sheath and penetrate the latter in order to insert themselves on 
the inner surface of the sheath. The openings in the notochordal 
sheath of Amphioxus, through which the ventral supporting fibres 
pass, were first observed by WILHELM MULLER in 1871. (W. 
MULLER, Ueber den Bau der Chorda dorsalis. Jenaische Zeit- 
schrift, VI. 1871. pp. 327-354.) See also PLATT (infra) and 
LWOFF (88). Latter contains complete bibliography of literature 
relating to structure of notochord. 

41 MAYER, PAUL. Uber die Entwicklung des Her sens und der 
grossen Gefdssstamme bei den Selachiern. Mitth. Zool. Stat. 
Neapel. VII. 1887. pp. 338-370. Taf. 11-12. 

42 MEYER, EDUARD. Studien liber den Korperbau der Anneliden. 
Mitth. Zool. Stat. Neapel. VII. 1887. pp. 592-741. Taf. 22-27. 

42 bis. MOREAU, CAMILLE. Recherches sur la Structure de la Corde 

dorsale de VAmphioxus. Bull. Acad. Belg. Tome 39. No. 3. 
1875. 22 pp. One plate. 

43 MULLER, WILHELM. Ueber die Stammesentiyicklwng des 
Sehorgans der IVirbeltliiere. 76 pp. Five plates. 4. Leipzig, 

44 MULLER, WILHELM. Ueber das Urogenitalsystem des Amphi- 
oxus und der Cydostomen. Jenaische Zeitschr. flir Naturwissen- 
schaft, Bd. II. (neu'e Folge). 1875. Sep. Abdruck. pp. 1-38. 
Two plates. 

This is the important work in which the pronephros and 
mesonephros were for the first time clearly distinguished from one 
another. The author was, however, in error regarding Johannes 
Mullers renal papillae of Amphioxus. 

45 MULLER, JOHANNES. Uber den Bau und die Lebenserscheinun- 
gen des Branchiostoma lubricum Costa, Amphioxus lanceolatus, 

Yarrell. Berlin, 1844. 4. 40 pp. Five plates. 
Read at the konigl. Akademie, 1841. 

46 NANSEN, FRIDTJOF. The Structure and Combination of the His- 
tological Elements of the Central Nervous System. Bergens 
Museums Aarsberetning for 1886. Bergen, 1887. 


47 OWSJANNIKOW, PHILIP. Ueber das Centralnervensystem des 
Ainphioxus lanceolatns. Bulletin de TAcad. imp. des Sciences de 
St. Petersbourg, Tome XII. 1868. pp. 287-302, with one plate. 
Also in Melanges Biologiques, T. VI. pp. 427-450. 

Introduced a method of maceration by which he was able to 
shake out the central nervous system and thus isolate it from the 
body. In this way he was able to correct the erroneous descrip- 
tions of de Quatrefages and others (who stated that there were 
ganglionic enlargements in the spinal cord), and to discover the 
alternate arrangement of the spinal nerves. 

48 PLATT, JULIA B. Fibres connecting the Central Nervous System 
and Chorda in Amphioxus. Anat. Anz. VII. 1892. pp. 282- 
284. Three figures in text. 

49 POLLARD, E.G. A New Sporozoon in Amphioxus. Quarterly 
Jour. Micro. Sc. XXXIV. N. S. 1893. pp. 311-316. Plate 

Unicellular parasites in intestinal epithelium. 

if) bis. POUCHET, GEORGES. On the Laminar Tissue of Amphioxus. 
Quarterly Jour. Micro. Sc. XX. N. S. pp. 421-430. Plate XXIX. 

50 DE QUATREFAGES, ARMAND. Memoire sur le systems nerveux 
et sur I" 1 histologie du Branchiostome ou Amphioxus. Annales des 
sciences nat. Zoologie. 3d series. IV. 1845. pp. 197-248. 
Plates 10-13. 

First observation of passage of ova through atriopore ; and 
discovery of the peripheral ganglion-cells in connexion with the 
cranial nerves. 

51 RATHKE, HEINRICH. Bemerkungen uber den Bau des Amphi- 
oxus lanceolatus, eines Fisches aus der Ordnung der Cy dost omen. 
Konigsberg, 1841. 4. pp. 1-38. One plate. 

52 RETZIUS, GUSTAV. Zur Kenntniss des centralen Nervensy stems 
von Amphioxus lanceolatus. Biologische Untersuchungen. Neue 
Folge II. pp. 29-46. Taf. XI. -XIV. Stockholm, 1890. 

52 bis. RETZIUS, GUSTAV. Das hintere Ende des Riickenmarks und 

sein Verhalten zur Chorda dorsalis bei Amphioxus lanceolatus. 
Verh. Biol. Vereins. (Biologiska Fb'reningens Forhandlingar.) 
Stockholm. Bd. IV. pp. 10-15. 9 n g s - ^gi- 

53 ROHDE, EMIL. Histologische Untersuchungen iiber das Nerven- 
system von Amphioxus lanceolatus. In Anton Schneider's Zoo- 
logische Beitrage. Bd. II., Heft 2. Breslau, 1888. pp. 169-211. 
Plates XV. -XVI. 

Standard work on the central nervous system of Amphioxus. 

54 ROHON, JOSEF VICTOR. Untersuchungen Yiber Amphioxus 
lanceolatus. Ein Beitrag zur vergleichenden Anatomie der Wir- 


belthiere. In Denkschriften der Math.-Naturwiss. Classe der kais. 
Akad. der Wissenschaften. Bd. XLV. Wien, 1882. 64 pp. 4. 
Six plates. 

Relates chiefly to nervous system. Describes also the smooth 
muscle-fibres in wall of pharynx, etc. Finds that the majority of 
sensory nerve-fibres to the skin end freely between the cells of the 
ectoderm in bush-like ramifications. For the rest, see NANSEN 

55 ROLPH, W. Untersuchungen uber den Ban, des Amphioxus 
lanceolatus. Morphologisches Jahrbuch, II. 1876. pp. 87-164. 
Taf. V.-VII. ; also figures in text. 

56 RUCKERT, JOHANNES. Entivickelung der Excretionsorgane. 
Ergebnisse der Anatomic und Entwicklungsgeschichte (Merkel 
und Bonnet), I. 1891. pp. 606-695. Includes an extensive bibli- 

57 SCHNEIDER, ANTON. Beitr'dge zur vergleichenden Anatomic 
und Entwicklungsgeschichte der Wirbelthiere. I. Amphioxus 
lanceolatus. pp. 3-31. Taf. XIV. -XVI. 4. Berlin, 1879. 

58 SCHULTZ, ALEXANDER. Zur Entwickelungsgeschichte des Sela- 
chiereies. Archiv. fur Mikr. Anat. XI. 1875. PP- 569-580. 
Taf. 34- 

Preliminary notes of both Semper and Schultz, regarding the 
segmental origin of the excretory tubules, were published in the 
Centralblatt fiir Medicinische Wissenschaft, 1874. 

59 SEMON, RICHARD. Studien ilber den Bauplan des Urogenital- 
sy stems der II 'irbelthiere ; dargelegt an der Entivickelung dieses 
Organsystems bei Ichthyophis glutinosus. Jenaische Zeitschrift, 
XXVI. 1891. pp. 89-203. Taf. I.-XIV. 

60 SPENGEL, J. W. Beitragzur Kenntniss der Kiemen des Amphi- 
oxus. Zool. Jahrblicher. Abth. fur Morphol. IV. 1890. pp. 257- 
296. Taf. 17-18. 

61 SPENGEL, J. W. Benhanfs Kritik meiner Angaben uber die 
Kiemen des Amphioxus. Anat. Anz. VIII. 1893. pp. 762-765. 

62 STIEDA, LUDWIG. Studien uber den Amphioxus lanceolatus. 
Mem. de TAcad. Imperiale des Sciences de St. Petersbourg, 7th 
series, Vol. XIX. No. 7. 70 pp. Four plates. 1873. 

Contains some good observations on the central nervous system. 
First to show that the split-like structure above central canal did 
not correspond to the posterior fissure of the vertebrate spinal cord, 
but was a portion of the original central canal itself, the lumen of 
which had been partially obliterated by approximation of its walls. 
First identification of ventral (motor) roots of spinal nerves in 


63 THACHER, JAMES K. Median and Paired Fins ; a Contribution 
to the History of Vertebrate Limbs. Transactions Connecticut 
Academy, III. No. 7. 1877. pp. 281-310. Plates 49-60. 

64 WEISS, F. ERNEST. Excretory Tubules in Amphioxus lanceolatus. 
Quarterly Jour, of Micro. Sc. XXXI. N. S. 1890. pp. 489-497. 
Plates 34-35- 

65 VAN WIJHE, J. W. Ueber Amphioxus. Anat. Anz. VIII. 1893. 
pp. 152-172. 

66 VAN WIJHE, J. W. Die Kopfregion der Cranioten beim Amphi- 
oxus, nebst B enter kungen fiber die ll'irbeltheorie des Sch'ddels. 
Anat. Anz. IV. 1889. pp. 558-566. 

67 VAN WIJHE, J. W. Ueber die Mesodermsegmente des Rumpfes 
und die Entivicklung des Excretionssy stems bei Selachiern. Archiv. 
f. Mikr. Anat. XXXIII. 1889. pp. 461-516. Taf. 30-32. 

68 WILLEY, ARTHUR. Report on a Collection of Amphioxus, made 
by Professor A. C. Haddon, in Torres Straits, 1888-89. Quarterly 
Jour. Micro. Sc. XXXV. N. S. January, 1894. pp. 361-371. One 
figure in text. 

Branchiostoma cultellum. Peters. 


69 AYERS, HOWARD. Bdellostoma Dombeyi, Lac. A Stiidy from 
the Hopkins Marine Laboratory. Biological Lectures, Marine 
Biological Laboratory, Woods Holl. 1893. No. VII. Boston, 

69 bis. BERT, PAUL. On the Anatomy and Physiology of Amphioxus. 
Annals and Mag. of Nat. Hist., 3d Series. Vol. XX. 1867. 
pp. 302-304. (Translated from Comptes Rendus. Aug. 26th, 
1867. pp. 364-367.) 

Breeding season of Amphioxus at Arcachon is from March to 
May. Was the first to observe the ejection of the sperm through 
the atriopore. Calls attention to remarkable lack of regenerative 
power in Amphioxus. Individuals cut in two will live for several 
days, but will not regenerate. " If the extremity of the body of 
an Amphioxus be cut off, the wound does not cicatrize ; on the 
contrary, the tissues become gradually disintegrated. I have 
seen animals, with only the tail mutilated, become gradually 
eaten away up to the middle of the branchial region, and live 
thus without any intestines, without abdominal walls, and without 
branchiae for several days." These observations of Paul Bert are 


capable of easy confirmation, and should be borne in mind in 
view of the extraordinary regenerative power which Wilson dis- 
covered in the segmentation stages of the embryo. 

70 BOVERI, THEODOR. Uber die Bildungsstatte der Geschlechts- 
tf n't sen und die Entstehiing der Genitalkammern bciin Atnphi- 
oxns. Anat. Anz. VII. 1892. pp. 170-81. Twelve figures. 

71 DOHRN, ANTON. Stitdien zitr Urgeschichte des U'irbelthier- 
korpers. III. Die Entstehiing und Bedeittung der Hypophysis 
bei Petroniyzon Planeri. Mitth. Zool. Stat. Neapel. IV. 1882. 

72 DOHRN, ANTON. Studien, VIII. Die Thyreoidea bei Petromy- 
zon, Amphioxus und 7^/nicafen. Ib. VI. 1885. 

Dohrn lays unnecessary stress upon the fact that often in 
transverse section, especially in the anterior region of the 
pharynx, the endostyle of Amphioxus projects up into the cavity 
of the pharynx in the form of a convex lens-shaped ridge. This 
is merely due to the muscular contraction of the pharynx, which 
almost invariably takes place when Amphioxus is placed in a 
killing reagent. It is, therefore, not an anatomical feature of 
any significance. 

73 DOHRN, ANTON. Studien, XII. Thyreoidea nnd Hypobran- 
chialrinne, Spritzlochsack nnd PseudobrancJiialrinne bei Fischen, 
Ammoccetes und Tunikaten. Ib. VII. 1887. 

74 DOHRN, ANTON. Studien, XIII. Uber Ner-ven und Gefdsse 
bei Ammoccetes und Petromyzon Planeri. Ib. VIII. 1888. 

75 FRORIEP, AUGUST. Entwickelungsgeschichte des Kopfes. 
Ergebnisse der Anat. und Entwickelungsgesch (Merkel und 
Bonnet), I. 1891. pp. 561-605. Eleven figures. 

Includes an extensive bibliography. 

76 HATSCHEK, BERTHOLD. Studien uber Entivicklung des Amphi- 
oxus. Arbeiten a. d. Zool. Institute. Wein, iSSi. 88 pp. 
Nine plates. 

77 HATSCHEK, BERTHOLD. Mittheilungen uber Amphioxus. 
Zoologischer Anzeiger, VII. 1884. pp. 517-520. 

Olfactory pit, sense-organ of praeoral pit, anterior preoral 
u nephridium." 

78 HATSCHEK, BERTHOLD. Uber den Schichtenbau von AmpJii- 
oxits. Anat. Anz. III. 1888. pp. 662-667. Five figures. 

Origin of sclerotome, etc. 

79 KASTSCHENKO, N. Zur Entwicklungsgeschichte des Selachier- 
embryos. Anat. Anz. III. 1888. pp. 445-467. 

One of the first to bring forward definite embryological facts to 
prove that the anterior (prae-auditory) head-cavities of VAN WIJHE 
(Ueberdie Mesodermsegmente, etc., des Selachierkopfes. Amster- 


dam, 1882) are not homodynamous with the true somites. He 
was followed in this respect by RABL (Theorie des Mesoderms. 
Morphologisches Jahrbuch, XV. 1889). 

80 KORSCHELT, E., und HEIDER, K. Lehrbuch der vergleichen- 
den Entwicklungsgeschichte der ivirbellosen Thiere. 3d Heft. 
Jena, 1893. 

8 1 KOWALEVSKY, ALEXANDER. Entwicklungsgeschichte des Am- 
phioxus lanceolatus. Mem. de 1'Acad. Imp. des Sciences de St. 
Pdtersbourg. VII. Series. T. XI. No. 4. 1867. Three 

82 KOWALEVSKY, ALEXANDER. Weitere Studien uber die Ent- 
wicklungsgescJiidite des Ampliioxus lanceolatus, nebst einem 
Beitrage zitr Homologie des Nervensy stems der Wurmer and 
Wirbelthiere. Arch. f. Mikr. Anat. XIII. 1877. pp. 181-204. 
Two plates. 

Among the definite discoveries communicated by Kowalevsky 
in these two memoirs may be mentioned the following : General 
features of segmentation and gastrulation, origin of mesoderm 
from archenteric pouches, unique method of formation of 
nerve-tube (see text), origin of notochord, neurenteric canal, 
asymmetrical origin of gill-slits and mouth, and in part the 

83 KUPFFER, CARL VON. Die Entivicklung von Petroinyzon 
Planeri. Arch. f. Mikr. Anat. XXXV. 1890. pp. 469-558. 
Six plates. 

Origin of head-cavities, hypophysis, etc. 

84 KUPFFER, CARL VON. Die Entwicklnng der Kopfnerven der 
Vertebraten. Verhandl. Anat. Gesellschaft in Miinchen. 1891. 
pp. 22-55. Eleven figures. (Erganzungsheft zum Anat. Anz. 
VI. 1891.) 

Ammocoetes (see Fig. 92 in text). 

85 KUPFFER, CARL VON. Studien zur vergleichende Entwick- 
lungsgeschichte des Kopfes der Kranioten I. Die Entwicklung 
des Kopfes von Acipenser sturio an Medianschnitten untersucht. 
pp. 95. Nine plates. Seven figures in text. Miinchen and 
Leipzig, 1893. 

Important contribution to the delimitation of the wall of the 
brain. On page 84 is a reconstruction of head-cavities of Am- 
mocoetes (see Fig. 72). Figs. 21 and 22 in the plates repre- 
sent cerebral vesicle of Amphioxus. (Cf. Fig. 51.) 

86 LANKESTER, E. RAY, and WILLEY. A. The Development of 
the Atrial Chamber of Amphioxus. Quarterly Jour. Micro. Sc. 
XXXI. 1890. pp. 445-466. Four plates. 


suchungen uber niedere Seethiere. Amphioxus lanceolatus. 
Miiller's Archiv f. Anat. u. Physiol. 1858. pp. 558-569. Taf. 

Description of larva; of Amphioxus taken off Heligoland. 
Drew attention to larval asymmetry, and to the existence of the 
brain-ventricle (cerebral vesicle). In absence of knowledge of 
early development their interpretation of many of the structures 
(especially prasoral pit, mouth, and gill-slits) was incorrect. 
Latter applies also to Schultze's observations. 

88 LWOFF, BASILIUS. Uber Bait und Entwicklttng der Chorda 
von Amphioxus. Mittheilungen a. d. Zool. Station. Neapel. 
IX. 1891. pp. 483-502. One plate. 

Consult this memoir for previous literature on histology of 

89 LWOFF, BASILIUS. Ueber einige ivichtige Punkte in der Ent- 
ivicklung des Amphioxus. Biologisches Centralblatt, XII. 1892. 
pp. 729-744. Eight figures. 

Notes absence of mesodermal " pole-cells." From frequency 
of mitoses in dorsal ectoderm of gastrula, concludes that the 
material destined to form dorsal wall of archenteron, from which 
notochord and myocoelomic pouches arise, grows in from the 
ectoderm round dorsal lip of blastopore. Hence notochord and 
mesoderm are essentially derived from ectoderm ! 

90 MARSHALL, A. MILNES. Vertebrate Embryology. London, 

91 MU'LLER, JOHANNES. Uber die Jugendzust'dnde einiger See- 
thiere. Monatsbericht der konigl. preuss. Akad. der Wissen- 
schaften zu Berlin. 1851. pp. 468-474. 

First accurate description of larva of Amphioxus, p. 474. In 
1847 Johannes Mliller obtained a young Amphioxus of 2i- mm. at 
Helsingfors. He says that the appearance of the gill-slits was 
peculiar, in that there were two rows of slits in the pharyngeal 
wall, placed one above the other. In the upper row were five 
round slits, while the lower slits were vertically elongated and 
were fourteen in number. He adds that it was doubtful whether 
it represented the young " Branchiostoma lubricum " or belonged 
to a new species. 

92 MULLER, WILHELM. Ueber die Hypobranchialrinne der Tuni- 
katen und deren Vorhandensein bei Amphioxus und den Cyklo- 
stomen. Jenaische Zeitschrift f. Naturvviss. VII. 1873. PP- 3 2 7~33 2 - 

93 PLATT, JULIA B. Further Contribution to the Morphology of 
the Vertebrate Head. Anat. Anz.VI. 1891. pp. 251-265. 


94 RABL, CARL. Uber die Mesoderms. Anat. 
Anz. III. 1888. pp. 667-673. Eight figures. 

Discovery of the sclerotome-diverticulum in embryo of Pristiurus. 

95 RICE, HENRY J. Observations upon the Habits, Structure, and 
Development of Amphioxns lanceolatus. American Nat. XIV. 
1880. pp. 171-210. Plates 14 and 15. 

Author was the first to find Amphioxus in Chesapeake Bay. 
With regard to development, he gives some fairly good figures of 
larva;, and observed some of the more obvious features of the 
metamorphosis, as already described by Kowalevsky. 

96 RUCKERT, JOHANNES. Ueber der Entstehung der Excretions- 
organe bei Selachiern. Arch, fur Anat. u. Physiol. (Anatomische 
Abtheilung). 1888. pp. 205-278. Three plates. 

Contains also the discovery of segmental origin of gonads. 

97 SCHNEIDER, ANTON. Beitrage zur vergleichenden Anatomic 
und Entivicklungsgeschichte der Wirbelthiere, II. Anatomic nnd 
Entwickl. von Petromyzon nnd Ammoc&tes. 4. Ten plates. 
Berlin, 1879. 

Figure of the ciliated grooves in pharynx of Ammocoetes, at 
page 84. 

98 SCHULTZE, MAX. Beobachtung junger Exemplar e von Amphi- 
oxus. Zeit. f. Wiss. Zool. III. 1851-2. pp. 416-419. 

Two larvae from Heligoland. Good description of structure of 

99 VAN WIJHE, J. W. Ueber Amphioxus. Anat. Anz. VIII. 
1893. pp. 152-172. 

100 WILLEY, A. On the Development of the Atrial Chamber of 
Amphioxus. (Preliminary communication.) Proceedings of the 
Royal Society, XLVIII. 1890. pp. 80-89. 

101 WILLEY, A. The Later Larval Development of Amphioxus. 
Quarterly Jour. Micro. Sc. XXXII. 1891. pp. 183-234. Three 

102 WILSON, EDMUND B. On Multiple and Partial Development 
in Amphioxus. Anat. Anz. VII. 1892. pp. 732-740. Eleven 
figures . 

In this and the following more detailed paper, the author 
describes and interprets a remarkable series of experiments on 
the artificial production of twins and dwarfs. Besides this, there 
are many important observations on the normal cleavage of the 


103 WILSON, EDMUND B. Amphioxus and the Mosaic Theory of 
Development. Journal of Morphology. VIII. 1893. pp. 579- 
638. Ten plates. 


104 ZIEGLER, H. ERNST. Der Ursprung der mesenchymatischen 
GewebebcidenSelachiern. Archiv f. Mikr. Anat. XXXII. 1888. 
pp. 378-400. One plate. 

Independent discovery of sclerotome-diverticulum. (See Rabl.) 


For bibliography relating to the Ascidians, see Professor W. A. HERD- 
MAX'S Reports on the Tunicata collected during the " Challenger" 
expedition Parts I. -III. 1882-88; and also KORSCHELT und 
HEIDER, " Lerhbuch der vergleichenden Entwicklungsgeschichte 
der wirbellosen Thiere." Heft III. Jena, 1893. 


105 AYERS, HOWARD. Concerning Vertebrate Cephalogenesis . 
Jour. Morph. IV. 1890-91. pp. 221-245. 

1 06 BATESON, WILLIAM. Memoirs on the Development of Balano- 
glossus. Quarterly Jour. Micro. Sc. Vols. XXIV. -XXVI. 


107 BROOKS, W. K. The Systematic Affinity of Salpa in its 
Relation to the Conditions of Primitive Pelagic Life ; the Phylogeny 
of the Timicata ; and the Ancestry of the Chordata. Part II. of 
Monograph of the Genus Salpa. Johns Hopkins University. 
Baltimore, 1893. 

1 08 BURCKHARDT, RUDOLF. Die Homologieen des Zwischenhirn- 
daches und ihre Bedeutitng fur die Morphologic des Hirns bei 
niederen Vertebraten. Anat. Anz. IX. 1894. pp. 152-155 and 

Relates to neuropore of craniate Vertebrates. Author calls the 
lobus olfactorius impar of Kupffer, the recessus neuroporicus. 

109 CLAPP, CORNELIA M. Some Points in the Development of the 
Toad-fish {Batr a chus Tait). Jour. Morph. V. 1891. pp. 494- 

Observations on the double origin of mouth, made in 1889, n t 
published in this paper. 

1 10 DAVIDOFF, M. VON. Ueber den " Canalis neurentericus 
anterior bei den Ascidien.'" Anat. Anz. VIII. 1893. pp. 301-303. 


in DOHRN, ANTON. Studien zur Urgeschichte des Wirbelthier- 
korpers, I. Der Mund der Knochenfische. Mitth. Zool. Stat. 
Neapel. III. 1881-2. pp. 253-263. 

112 FIELD, GEORGE W. The Larva of Asterias vulgaris. 
Quarterly Jour. Micro. Sc. XXXIV. 1892. pp. 105-128. 

113 FOWLER, G. HERBERT. The Morphology of Rhabdopleura 
Normani Allman . Festschrift fur Rudolf Leuckart. pp. 293-297. 
Leipzig, 1892. 

114 HARMER, S. F. See M'INTOSH. 

115 HERDMAN, W. A. Article " Tunicata." Ency. Brit, gth ed., 
republished in "Zoological Articles" by Lankester, etc. 

116 HUBRECHT, A. A. W. Article " Nemertmes ." Ency. Brit, 
gth ed., republished in "Zoological Articles" by Lankester, etc. 

n6bis. HUBRECHT, A. A. W. On the Ancestral Form of the 
Chordata. Quarterly Jour. Micro. Sc. XXIII. 1883. pp. 349-368. 
For later works on this subject see Notes to Chap. V. 

117 KUPFFER, C. VON. Eittwickelungsgeschichte des Kopfes. In 
Merkel and Bonnet's Ergebnisse der Anatomic und Entwickelungs- 
geschichte, II. 1893. pp. 501-564. 

118 LANG, ARNOLD. Znm Verst'dndnis der Organisation von 
CepJialodiscus dodecaloplius Afhit. Jenaische Zeitschrift f. 
Naturwiss. XXV. 1891. 

119 LANG, ARNOLD. Ueber den Einfluss der festsitzenden Lebens- 
weise auf die Thiere. Jena, 1888. 

1 20 LANKESTER, E. RAY. Degeneration : a Chapter in Darwinism. 
Nature Series. London, 1880. Republished in ' The Advance- 
ment of Science ; Occasional Essays and Addresses." London, 

121 LANKESTER, E. RAY. A Contribution to the Knowledge of 
RJiabdopleura. Quarterly Jour. Micro. Sc. XXIV. 1884. pp. 

122 MACBRIDE, E. W. The Organogeny of Asterina Gibbosa. 
Proceedings Royal Society. Vol. 54. 1893. pp. 431-436. 

123 M'lNTOSH, WILLIAM C. Report on Cephalodiscns dodecalo- 
phus, APIntosh. " Challenger" Reports. Zoology, XX. 1887. 
With Appendix by S. F. HARMER. 

124 MORGAN, T. H. The GrowtJi and Metamorphosis of Tornaria. 
Jour. Morph. V. 1891. pp. 407-458. 

125 MORGAN, T. H. The Development of Balanoglossus. Jour. 
Morph. IX. 1894. pp. 1-86. 

126 PLATT, JULIA B. Further Contribution to the Morphology of 
the Vertebrate Head. Anat. Anz. VI. 1891. pp. 251-265. 

Describes the double origin of mouth in Batrachus. 


127 POLLARD, H. B. Observations on the Development of the Head 
in Gobius capita. Quarterly Jour. Micro. Sc. XXXV. 1894. 

PP. 335-352. 

127 bis. POLLARD, H. B. The " Cirrhostomial" Origin of the Head 

in Vertebrates. Anat. Anz. IX. 1894. pp. 349-359. 

128 RABL-RUCKHARD, H. Der Lobus Olfactorius Impar der 
Selachier. Anat. Anz. VIII. 1893. pp. 728-731. 

129 SEDGWICK, ADAM. The Original Function of the Canal of the 
Central Nervous System of Vertebrata. Studies from Morph. 
Lab. Cambridge, II. 1884. pp. 160-164. 

130 SEDGWICK., ADAM. Notes on Elasmobranch Development. 
Quarterly Jour. Micro. Sc. XXXIII. 1891-92. pp. 559-586. 

Contains important observations on the first appearance of the 
mouth, and its relation to the pituitary body. 

131 SEELIGER, OSWALD. Studien zur Entwicklungsgeschichte der 
Crinoiden. {Antedon rosacea.} Zoologische Jahrbiicher. Abth. 
f. Anat. VI. 1892. pp. 161-444. 

132 VAN WIJHE, J. W. Ueber den vorderen Neuroporus und die 
phylogenetische Function des Canalis Neurentericus der Wirbel- 
thiere. Zool. Anz. VII. 1884. pp. 683-687. 

133 WILLEY, A. Studies on the Protochordata, /.-///. Quarterly 
Jour. Micro. Sc. XXXIV.-XXXV. 1893. 

Contain further bibliographical references. 


Acipenser sturio, 102, 129, 287. 

Acrania, 17, 46. 

AGASSIZ, A., 250, 251, 256. 

ALLMAN, 262. 

Ammoccetes, 163-170, 173, 178, 182, 186, 


ANDREWS, 39, 41. 
Annelid theory, 5, 79, 82, 97, 176, 282, 

290, 293. 
Annelids, excretory system of, 78-82, 99. 

giant fibres of, 97, 103. 

nervous system of, 95-97. 

segmentation of, 4. 

vascular system of, 55. 
Antedon rosacea, 256, 268-269, 271. 
Anus, 14, 25, 118, 131, 187. 
Aorta, dorsal, 49, 50, 53. 
Aperture, buccal, 182. 

cloacal, 182, 183, 210. 
Appendicularia, 180, 236-239, 241, 277. 
Archenteron, no. 
Artery, branchial, 47, 50, 98, 139. 

genital, 98. 
Ascidians, pelagic, 181, 236. 

sessile, 181. 

Asterias vulgaris, 254, 270. 
Asterina gibbosa, 270, 271. 
Asymmetron lucayanum, 40, 41. 
Asymmetry, 155-162, 177. 
Atriopore, 14, 77, 105. 
Atrium (see also Cavity, peribranchial), 
14, 22, 186, 195. 

development of, 75-78, 210-212. 

post-atrioporal extension of, 25. 
Audition, 44. 
AUDOUIN, 197. 

Auricularia, 251-253, 256, 268. 
Axis (see Relations, axial). 
AYERS, 18, 173. 

Balancers, 42. 

Balanoglossus, 29, 43, 98, 128, 221, 222, 

231, 242-253, 259, 261, 264, 265, 

274, 276. 

Balanoglossus, nervous system of, 244- 

Kowalevskii, 248, 250. 

Kupfferi, 248, 253. 
BALFOUR, 5, 38, 79, 175, 190, 203, 273, 

283, 292. 
Band, adoral ciliated, 250. 

circumoral ciliated, 251, 256. 

longitudinal ciliated, 251. 

post-oral (circular) ciliated, 251, 256. 
Bands, mesodermic, 120, 217, 218. 

peripharyngeal, 34, 140, 145, 168-169, 

179, 185, 195, 226. 
Bars, branchial (see Gill-bars). 
BATESON, 98, 221, 244, 245, 250, 259, 

263, 291. 

Batrachus tau, 281. 
Bdellostoma, 173, 285. 
BEARD, 208, 281, 292. 
VAN BENEDEN, 187, 191, 197, 200, 224, 


BENHAM, 33, 42. 
BERT, 174. 
Bipinnaria, 251. 
Blastocoel, 108, 254, 255. 
Blastomeres, 107. 
Blastopore, no, 112, 197. 
Blastula, 108, 197. 
Blood-sinuses, 191, 192. 
Blood-vessels, contractile, 47, 98. 

origin of, 122. 
Bodies, polar, 106. 
Body, pineal, 207. 

pitituary (see Hypophysis). 
Body-cavity (see also Coelom), 217, 220- 
222, 247. 

prceoral, 128, 218. 
Bojanus, organ of, 194. 
Botryllus, 181, 240. 

BOURNE, A. G., 81. 

BOVERI, 42, 48, 60, 98, 99, loo, 151, 177. 

Brachiolarui, 270. 



Brain, 92, 101. 
Branchiomery, 65, 132. 
Branchiostoma cultellum, 40. 

lubricum, 8. 
Breeding-season, 105. 
Brood-pouch, 215. 
BROOKS, 254, 277, 289. 
Bulbils, vascular, 48. 


BURY, H., 269. 


Canal, alimentary, 24, ill, 187, 196, 214, 
235, 249, 264. 

neurenteric, 114, 118, 199, 202, 275. 
Capillaries, 49, 98. 
Cap it f 1 1 ides, 81. 
Cartilages, buccal, 18, 147. 

labial, 18. 
Caulus, 266. 
Cavity, opercular, 22. 

peribranchial (see also Atrium), 22, 
183, 186, 195, 209. 

peritoneal, 22. 

Cells, epithelio-muscular, 191. 
Cellulose, 182. 
Cenogenesis, 177. 
Cephalisation, 75, 89. 
Cephalochorda, 13. 
Cephalo discus, 261-267, 2 %> 2 %9- 
ChcEtognatha, 278. 

Ciona intestinalis , 203, 210, 215, 222, 224, 
226, 229, 230-235, 240, 271, 288, 
292, 293. 

Cirri, buccal, 12, 20, 145. 
Cladoselachidce, 44. 
Clavelina, 181, 185, 187, 200, 215, 225, 

241, 288. 
Cleavage, 107, 197. 

polymorphic, 108. 
Cceca, intestinal, 249, 261. 
Casciliani, 67. 
Ccecum, hepatic, 24, 236. 
Ccelom, 22, 26, 31, 33, in, 121, 122, 220- 
222, 247-248, 265, 266. 

perigonadial, 153, 177. 
Ccenoecium, 263. 
Collar-pores, 98, 248, 265. 
Collar-region, 242, 264. 
Collector, 45, 165. 
Commissure, circumoesophageal, 96, 273, 

Compression, bilateral, 15, 43, 115. 

Contraction, peristaltic, 98, 192. 

Cordon ganglionnaire visceral, 224. 

COSTA, 7, 10. 

Craniota, 17. 

Crinoidea, 268. 

Cross-bars, 28. 


Cutis, 38, 41, 122. 


Cyclostomata, 8, 10, 45, 208. 

Cyclostome, 46. 

Cynthia papillosa, 200. 


DEAN, B., 44. 

Degeneration, 5. 

Development, abbreviated, 214, 215, 239. 

adolescent period of, 149, 150. 

direct, 250. 

duration of larval, 149, 169, 203, 215. 

embryonic, 114, 201. 

larval, 117, 130. 

latent, 145, 160. 

precocious, 161, 212. 
Differentiation, sexual, 154. 
Dissepiments (see Septa). 
Distaplia magnilarva, 206, 225, 288. 
Distribution, n, 40-41. 
Diverticula, anterior intestinal (see also 

Head-cavities), 115. 
DOHRN, 5, 30, 167, 173, 176, 178, 179, 280, 

Duct, mesonephric, 66. 

pronephric, 69, 78, 99. 
Dura mater, 87. 

Echinoderms, 250-256, 267-271, 291. 
Ectoderm, 24, 78. 

ciliated, 112, 113, 130, 175, 243, 257. 

definitive, in. 

primitive, no. 
EISIG, 45, 81, 94, 103, 293. 
Embryo, ciliated, 113, 214. 

ventral curvature of Ascidian, 201. 
EMERY, 67. 
Endoderm, definitive, HI. 

primitive, no. 
Endostyle, 9, 24,31, 39, 130, 138, 149, 150, 

167, 177, 185, 195, 227, 229, 250. 
Enteroccel, 252, 254, 255. 
Enteropneusta, 242. 
Epicoele, 41. 
Epithelium, atrial, 33, 59, 100, 209. 

ccelomic, 33, 122, 220-222. 



Equilibration, 44, 205. 
Equilibrium, 10, 43. 
Evolution, parallel, 80, 247, 290. 
Eye of Ascidian tadpole, 102, 206. 
Eye, median, 18, 102, 130. 

myelonic, 207. 

pineal, 207-209. 
Eyes, paired, 102. 

Fascia, 36, 123. 
FELIX, 99. 

Fertilisation, 106, 188. 
Fibres, giant, 92-94, 103. 

Miillerian, 94. 

of Mauthner, 94. 

supporting, 89. 
FIELD, G. W., 254. 

Fierasfer, 67. 

Fin, definitive caudal, 131. 

provisional caudal, 115. 
Fin-rays, 15. 
Fins, 15, 44. 

lateral, 38, 42. 

Fixation, organ of, 222, 229, 271, 280. 
Flexure, cranial, 92, 279. 
FOL, 239. 
Folds, medullary, 199. 

metapleural, 15, 38, 42, 43, 76, 132, 176. 
Follicle, 105. 
Food, 9, 39, 185, 249. 
FOWLER, G. H., 262, 266, 267. 
FRORIEP, 175. 

Function, change of, 176, 280. 
Funnels, atrio-ccelomic, 58, 98. 

brown (same as preceding). 

ccelomic (see also Nephrostomes), 

FUSARI. 87, 163. 

Fusari, plexus of, 87, 178. 


Ganglia, peripheral, 85, 88. 

spinal, 84, 103. 
Ganglion, Ascidian, 188, 224, 225. 

cerebral, 96, 270, 272-274. 
Ganglion-cells, 89, 91. 

bipolar, 95. 

giant, 92. 

multipolar, 92. 
GARSTANG. 240, 250. 
Gastrula, no, 197. 

significance of, ill. 

Gastrulation, 109. 
GEGENBAUR, 249, 273. 
Germ-layers, primitive, no, 114. 
Gill-bars, 28, 32-34. 

blood-vessels of, 48-49. 
Gill-pouches, 165, 166. 
Gill-slit, first, 117, 118, 132, 141, 166, 170- 


Gill-slits (see also Stigmata), 17, 27, 100, 
130-132, 135-138, 139, 148-149, 1 60, 
173-174, 195, 229, 234, 243, 244, 264, 

asymmetry of, 157-158. 
atrophy of, 140, 143, 149. 
Gland, club-shaped, 116, 117, 134, 138, 

141, 170-172, 176. 
pyloric, 236. 
subneural, 188-191, 225. 
thyroid, 169-170. 
thymus, 29, 30. 
Glands, fixing, 204. 
Glomerulus, 64, 65, 69, 100. 
Gnathostome, 46. 
Gobius capita, 282. 
DE GRAAF, 208. 

Groove of Hatschek, 21, 51, 135. 
Groove, epibranchial, 226. 
hyperbranchial, 34, 39, 195. 
hyperpharyngeal (same as preceding) . 
hypobranchial (see also Endostyle), 

9, 167. 

medullary, 112, 198. 
pericoronal (see Bands, peripharyn- 

peripharyngeal (see Bands, peripha- 

ryngeal) . 
Gut, post-anal, 203, 

HAECKEL, 5, 46, in, 177. 

HANCOCK, 190. 

HARMER, 263, 289. 


HATSCHEK, 41, 91, 102, 103, 104, 112, 115, 

118, 174, 175, 292. 
Hatschek's nephridium, 172. 
Head-cavities of Ammoccetes, 129. 

of Amphioxus, 126-128. 

praemandibular, 128, 175, 279-280. 

of Sagitta, 277. 
Heart, 46, 51-53, 191, 192. 

recurrent action of, 193. 
Heptanchus, 173. 


HERDMAN, 183, 277, 293. 
Hermaphrodite, 187, 196. 
Hcxanchus, 173. 
HjORT, 225, 293. 


Hood, nerve-plexus of oral, 84, 178. 
oral, 12, 147, 150, 178. 

HUBRECHT, 258, 259, 260, 287, 291. 

HUXLEY, 20, 22, 41, m. 
Hypophysis, 160, 165, 178, 190, 191, 195, 
225, 283-288, 290, 292. 

Ichthyophis glutinosus, 67. 
Infundibulum, 102, 283, 285. 
Insects, compared with Vertebrates, 2-4. 
Involutions, atrial, 209, 241. 

JULIN, 187, 190, 197, 200, 224, 225, 226, 


Kidney, 65. 


KQlliker's olfactory pit, 19. 

KOPPEN, 103. 

KORSCHEl.T and HEIDER, 178. 

KOVVALEVSKY, 4, 104, 114, 174, 196, 2l6, 

KROHN, 197, 250. 

K.UPFFER, 101, 102, 128, 129, 175,283,287. 

Lamella, post-oral, 264. 

Lamina, dorsal, 183, 185, 195, 226. 

terminalis, 284. 
Lamprey (see Petromyzon). 
LANG, 291. 

LANGERHANS, 21, 56, 98, 101, 154. 
Lanice conchilega, 80. 
LANKESTER, 38, 41, 58, 62, 98, in, 237, 

262, 266. 

Ligamentum denticulatum, 25,63, 164. 
Limax lanceolatus, 7. 
Line, lateral, 21, 42-45. 
Liver, 24. 

Lobe, prpeoral, 218, 222, 228, 229, 254, 
267-280, 290, 292. 

procephalic, 272. 

Lobus olfactorius impar, 102, 283, 284. 
Locomotion, caudal, 103, 203. 

ciliary, 121. 

muscular, 121. 
Loimia medusa, 80. 

Lumbricus, 79, 272. 
LWOFF, 175. 
Lymph-spaces, 15, 51. 

Mantle, cellulose, 183. 

muscular, 183. 
Maturation, 106. 
Mauthner, fibres of, 94. 
MAYER, PAUL, 99, 100. 
Medulla oblongata, 91. 
Membrane, interccelic, 152. 

vitelline, 105. 
Merlucius, 67. 

Mesenchyme, 201, 217, 220-222, 261. 
Mesoderm, in, 114, 120, 122, 199-201, 


Mesonephros, 66. 

Metamerism, 64, 132, 196, 246-247, 291. 
Metamorphosis, 136, 150, 215, 223, 250, 256. 
Metanephros, 66. 
METCALF, 293. 
MINOT, 155. 

M'lNTOSH, 263. 

Molgula, 194. 

Molgula manhattensis, 210, 232, 240. 

MORGAN, T. H., 232, 245, 247, 253, 256, 

Mouth, 19, 117, 131, 143-144, 146, 150, 

176, 178, 229, 276, 280-282. 
asymmetry of, 157-160. 
MULLER, J., 8, 18, 50, 56, 59, 250. 
MULLER, W., 102, 167. 
Muscles, 34-37, S6, 122, 195, 203, 222, 


Muscle-fibres, origin of, 121. 
Musculature (see Muscles). 
Myocoel, 121. 
Myotomes, 13, 150. 
AJyxine, gill-slits of, 171. 

hypophysis of, 285. 

pronephric duct of, 100. 

NANSEN, 103. 


Nemertines, 249, 256-261, 272, 273. 

lateral nerves of, 259. 

medullary nerve of, 259, 260. 
Mephridium, 62, 79, 99, 261. 



Nephrostomes, 65, 69, 72. 
Nerve-cord, ventral, 96, 259, 273, 289. 
Nerves, cranial, 85. 
motor, 86, 100. 
R. branchialis vagi, 163, 164. 
Rr. cutanei ventrales, 44. 
R. recurrens trigemini et facialis, 45. 
R. cutaneus qumti (same as preced- 

R. lateralis trigemini (same as pre- 

R. dorsalis, 85, 103. 
R. lateralis vagi, 45, 259. 
R. ventralis, 85, 103. 
R. visceralis, 86. 
sensory, 86. 
spinal, 83. 

Nerve-tube (see Tube, medullary). 
Nervous system, origin of central, in, 

119, 198. 
Neuropore, 19, 90, 115, 160, 199, 202, 

223, 225, 283, 285, 287, 292. 
NotidanidcB, 173. 

Notochord, 8, 13, in, 115, 124-126, 158, 
161-162, 199, 216, 222, 244, 266, 
286, 287, 290. 

Ontogeny, 177. 
Operculum, 264. 
Organs, renal, 55, 194. 

reproductive (see also Pouches, 
gonadic), 122, 151-155, 187-188, 
246, 266. 
Otocyst, 205. 
Otolith, 10, 205, 224. 
Oviduct, 187. 
Ovum, 105. 

Palingenesis, 177. 

Paludlna vivipara, 220. 
Papillae, adhesive, 204. 

renal, 56-57, 59. 
Pericardium, 191, 218. 
PerichcEta, 81. 

Petromyzon, 93, 163, 169, 286. 
Phallusia, 203, 232, 292. 
Pharynx, 27, 183. 
Phylogeny, 177. 

Pigment, 18, 26, 33, 102, 130, 131, 134, 

Pigment-cells, 135. 
Pilidium, 272. 

Pit, olfactory, 19, 90, 145, 160, 165, 195, 
283, 285, 292. 

prasoral, 51, 128, 135, 144, 148, 267. 
Plate, apical, 255-256, 269, 270, 272-274, 

medullary, 113, 115, 118, 198. 
Plates, skeletal (endostylar), 32. 
Pleuronectidce, 3, 40, 162, 178. 
Plexus, branchial, 163, 164, 165. 
Pluteus, 268, 270. 
Pole-cells, mesoblastic, 175. 
POLLARD, H. B., 282. 
Pontob delta, 81. 
Porus branchialis, 23. 
Pouches, archenteric, 114, 115, 120, 247, 

gonadic, 13, 25, 40, 153-154. 

myocoelomic, 122. 
POUCH ET, 82. 
Pristiurtts, 99. 

Proboscis, 221, 242, 247, 257, 264. 
Proboscis-cavity, 247. 
Proboscis-pore, 128, 248, 253, 264. 
Proboscis-sheath, 258. 
Products, genital, 174. 
Pronephros, 66-75, 7&- 

blood-vessels of, 63, 69, 74, 100. 

development of, 69, 78. 
Prostomium, 272. 
Protopterus, 14. 
Pyrosonta, 181, 236, 241. 

QUATREFAGES, 88, 174. 

RABL, 175. 

Raderorgan, 21, 148. 
Recessus opticus, 102. 

.ectus abdojninis,35; 

r, axial, 226-229. 

.ETZIUS, 82, loo, 103. 
Rhabdopleura, 261, 262, 266, 267. 
Ridge, epibranchial, 226. 
Ridges, subatrial, 76. 
RITTER, 250. 
Rods, skeletal, 28. 

ROHDE, IOO, IOI, 103. 

ROHON, 82, 86, 163, 165. 
ROLPH, 23, 41, 56, 86, 98. 
RUCKERT, 60, 99, ioo, 154. 



Sac, branchial (see also Pharynx), 183, 

195, 227. 

Sagitta, 13, 277-278. 

principles of, 2, 279. 
Salpa, 180, 182, 193, 236, 241. 
Sarcolemma, 36. 
SARS, G. O., 262. 
SAVIGNY, 190. 
Schizocoel, 175. 
SCHNEIDER, ANTON, 35, 38, 98, 100, 178. 
Sclerotome, 123, 175, 221. 
SEDGWICK, ADAM, 112, 289, 291. 
SEELIGER, 239-240, 269, 277. 
Segmentation (see Cleavage). 
Segmentation-cavity, 108. 
SEMON, 67. 

SEMPER, 5, 79, 99, 176. 
Sense-cells, 20, 21. 
Sense-organ of prasoral pit (see Groove 

of Hatschek). 
Septa, 13, 37, 122. 
Sheath, notochordal, 38, 123. 
Shield, buccal, 263. 
Skeleton, axial, 13. 
Snout, 115, 218. 
Somites, mesodermic, 115, 121. 
Spawning, 105. 
Species of Amphioxus, 41. 
SPEE, GRAF, 99. 

SPENCER, BALDWIN, 207, 208, 209. 
SPENGEL, 38, 41, 248. 
Spermatozoa, 105. 
Spinal cord, 83, 222. 

central canal of, 89, 289. 
Spiracle, 173. 
Spiraculum, 23. 
Splanchnocoel, 122. 
Stage, critical, 149, 174. 
STIEDA, 100. 
Stigmata, 183, 195, 196, 227. 

formation of, 229-235. 
Stomodoeum, 165, 209. 
Sympathetic system, 35, 86. 
Synapticula (see Cross-bars). /Q 

Table, showing order of development of 
Ascidian and Amphioxus, 213. 

Tadpole, Batrachian, 14. 

Tail of Ascidian tadpole, 201-204, 212, 


Teleosteans, 45, 281. 
Tentacles, velar, 20, 195. 
Test, 182, 240. 
Thymus, 29. 
Tissue, connective, 37, 41, 122. 

mesenchymatous, 221. 
Tongue-bars, 28, 140, 142, 148, 231. 
Tornaria, 250-253, 255-256, 270, 274. 
Trochophore, 256, 272. 
Tube, medullary, 114, 120, 198, 274. 

neuro-hypophysial, 225. 
Tubercle, dorsal, 189, 225. 
Tuberculum posterius, 102. 
Tubules, excretory, 59-65, 72, 100, 122. 

mesonephric, 70, 177. 

pronephric, 67, 70, 78, 100. 

uriniferous, 65. 
Tunic (see Test). 

Ureter, 66. 
Urmund, no. 
USSOW, 190. 

Vacuolisation of notochord, 125, 216, 

240, 244. 

Vas deferens, 187. 
Vein, cardinal, 54. 

caudal, 54. 

hepatic, 49, 98. 

portal, 53, 98. 

sub-intestinal, 49, 53-55. 
Velum, 20, 50, 150, 178. 
Vesicle, cerebral, 90, 100, 204, 223, 224, 

Water-pore, 253, 254. 

WEISS, F. E., 57, 59. 

VAN WIJHE, 39, 50, 51, 88,99, la8 . l6 3. 

164, 165, 178, 289. 
WILSON, E. B., 108, 174, 175, 293. 
WOODWARD, A. S., 44. 


ZIEGLER, H. E., 175. 
Zoarces, 67.